ARTHRITIS & RHEUMATOLOGY
Vol. 67, No. 11, November 2015, pp 2837–2844
DOI 10.1002/art.39297
C 2015, American College of Rheumatology
V
Autoantibodies in Prediction of the Development of
Rheumatoid Arthritis Among Healthy Relatives
of Patients With the Disease
Cesar Ramos-Remus,1 Jose Dionisio Castillo-Ortiz,2 Luis Aguilar-Lozano,2
Jorge Padilla-Ibarra,2 Carlos Sandoval-Castro,2 Cesar Omar Vargas-Serafin,2
Hector de la Mora-Molina,2 Ariadna Ramos-Gomez,2 Adriana Sanchez-Ortiz,2
Hilario Avila-Armengol,3 and Francisco Javier Aceves-Avila4
Objective. Although blood bank–based studies have
shown that rheumatoid arthritis (RA)–related autoanti-
bodies are present before the onset of RA, information
on their positive predictive value (PPV) for development
of RA in healthy individuals is scarce. This study was
undertaken to assess the 5-year PPV of serum IgM rheu-
matoid factor (IgM-RF) and anti–cyclic citrullinated
peptide (anti-CCP) for the development of RA among
healthy relatives of patients with RA.
Methods. Healthy relatives of RA patients were
invited to participate in a cohort study. At baseline, infor-
mation on participants’ medical history was obtained, and
serum levels of IgM-RF and anti-CCP antibodies were
determined (by nephelometry and second-generation anti-
CCP enzyme-linked immunosorbent assay, respectively).
The subjects were followed up every 4 months via a
Funds from an unrestricted grant from Pfizer (0881A-
102383) were used for erythrocyte sedimentation rate, IgM rheuma-
toid factor, and anticyclic citrullinated peptide testing and to obtain
office supplies for the present study.
1
Cesar Ramos-Remus, MD, MSc: Centro de Investigacion
Biomedica de Occidente, Instituto Mexicano del Seguro Social and
Unidad de Investigacion en Enfermedades Cronico-Degenerativas,
Guadalajara, Mexico; 2Jose Dionisio Castillo-Ortiz, MD, Luis Aguilar-
Lozano, MD, Jorge Padilla-Ibarra, MD, Carlos Sandoval-Castro, MD,
Cesar Omar Vargas-Serafin, MD, Hector de la Mora-Molina, MD,
Ariadna Ramos-Gomez, Adriana Sanchez-Ortiz, MD: Unidad de
Investigacion en Enfermedades Cronico-Degenerativas, Guadalajara,
Mexico; 3Hilario Avila-Armengol, MD: Hospital Civil Dr. Juan I.
Menchaca, Guadalajara, Mexico; 4Francisco Javier Aceves-Avila, MD,
PhD: Hospital General Regional 46, Instituto Mexicano del Seguro Social,
Guadalajara, Mexico.
Address correspondence to Cesar Ramos-Remus, MD, MSc,
Unidad de Investigacion en Enfermedades Cronico-Degenerativas, Colo-
mos 2292, Providencia Guadalajara, 44620, Mexico. E-mail: rramos@
cencar.udg.mx.
Submitted for publication March 31, 2015; accepted in revised
form July 21, 2015.
2837
structured interview (Community Oriented Program for
Control of Rheumatic Diseases [COPCORD] question-
naire). When the COPCORD questionnaire indicated
possible arthritis, subjects underwent an in-office rheu-
matology assessment including joint count. The study
end point was defined as fulfillment of the American
College of Rheumatology criteria for RA.
Results. Eight hundred nineteen initially healthy
relatives of 252 patients with RA were included (69%
female, 41% offspring, mean 6 SD age 35 6 12 years).
Eleven (1.3%) were positive for both anti–CCP-2 and RF,
12 (1.5%) only for anti–CCP-2, and 16 (2%) only for RF.
RA developed in 17 (2.1%) of the relatives during the
5-year followup (3,313 person-years for the seronegative
group and 60.8 person-years for the anti–CCP-2–positive
group). The PPV was 64% when both anti–CCP-2 and
RF were positive and 58% when only anti–CCP-2 was
positive. Offspring of patients with RA had an indepen-
dent 3-fold increased risk of developing RA.
Conclusion. Results of the present study indicate
that the magnitude of risk for developing RA in healthy
relatives of patients with RA can be estimated using
simple routine laboratory tests.
Rheumatoid arthritis (RA) has multidimensional
effects, ranging from pain and progressive destruction
of diarthrodial joints to development of comorbid con-
ditions, family distress, and high societal costs (1–3).
Although the incidence of RA peaks during the fifth
decade of life, in some countries the disease begins
before the age of 35 years in almost 50% of cases (4).
Early initiation of an optimal treatment strategy
can limit the overall impact of RA (5,6), including pre-
venting damage and allowing patients to continue to
2838
work (7,8). This has led to the concept of the “window of
opportunity” for treatment (9). However, treatment dur-
ing this window of opportunity does not always happen,
for reasons including patient delay in seeking medical
attention, time taken by the primary care provider to
diagnose the disease and refer the patient to a rheuma-
tologist, and wait time to obtain an appointment with a
rheumatologist (10–12), among other factors. Therefore,
any means of detecting RA at very early stage, or even
during the preclinical phase of the disease, would have
meaningful benefit for the patient.
The prespecified working hypothesis of the pre-
sent study was that it is possible to identify, with a good
degree of accuracy and using laboratory tests (13) com-
monly available in clinical settings, those healthy rela-
tives of RA patients who will later develop RA. This
hypothesis is based on the following evidence: 1) positiv-
ity for rheumatoid factor (RF) and/or anti–cyclic citrulli-
nated peptide (anti-CCP) can precede the clinical
manifestation of RA by several years (14–22); 2) based
on studies of Caucasian populations, the genetic contri-
bution to RA has been estimated at 12–60% of the vari-
ation in disease liability (23,24); and 3) healthy relatives
of RA patients have an increased risk of developing the
disease (25). Therefore, healthy relatives of patients with
RA would be an appropriate population in which to inves-
tigate outcome predictors for the development of the dis-
ease, in order to decrease delay in the diagnosis and to
increase the proportion of patients who receive treatment
during the window of opportunity. Consequently, the aim
of this study was to assess the 5-year predictive value of
serum IgM-RF and anti-CCP in the development of RA
in healthy relatives of patients with RA.
SUBJECTS AND METHODS
This was a prospective cohort study of healthy individ-
uals older than 15 years who had no history of chronic dis-
eases, including arthritis or any other rheumatic disease, and
who had a relative with a diagnosis of RA. A flow chart
describing the enrollment and followup protocol is available
upon request from the corresponding author.
RA patients. The study was explained to consecutive
patients with a diagnosis of RA who attended the outpatient
rheumatology clinic at 1 of 3 hospitals (1 tertiary care, 1 second-
ary care, 1 university-based) in Guadalajara, Mexico between
January and March 2007. Patients were asked to invite first-
and second-degree relatives (related by blood, not adoptive or
related only by marriage) not having RA or any other rheumat-
ic disease to participate in the study. Patients were instructed to
make an appointment for themselves (with a referral letter
from their rheumatologist) and their relatives at a research cen-
ter. At the research center the patients’ clinical history was
recorded and joint assessment was performed by 2 certified
clinical rheumatologists, to confirm the RA diagnosis according
RAMOS-REMUS ET AL
to classification criteria at that time (the American College of
Rheumatology [ACR] 1987 criteria [26]). On the same day, IgM-
RF was measured by nephelometry, and second-generation anti-
CCP IgG by enzyme-linked immunosorbent assay (ELISA)
(Architect i1000SR; Abbott Diagnostics). IgM-RF was consid-
ered positive if $20 IU/ml, and anti–CCP-2 if $7 IU/ml.
Relatives. Clinical history was also ascertained for the
relatives brought by RA patients to the research center. The
Community Oriented Program for Control of Rheumatic Dis-
ease (COPCORD) questionnaire (27) was administered and
joint assessment was performed to determine the absence of
any current or past relevant diseases, including arthritis or oth-
er rheumatic diseases. Relatives were asked to confirm that
they were in fact biologically related to the RA patient; in-laws
or adoptive relatives were excluded. Erythrocyte-sedimentation
rate (ESR), IgM-RF, and anti–CCP-2 were measured using the
same testing procedures as in RA patients.
The COPCORD instrument is a 5-item questionnaire
designed for population-based screening to detect rheumatic
and allied musculoskeletal disorders, which has been translat-
ed into Spanish. Relatives were classified as healthy if they had
no relevant medical history, normal results on physical exami-
nation including joint assessment, “negative” results on the
COPCORD questionnaire (i.e., no indication of a rheumatic
or musculoskeletal disease), and a normal ESR. The cohort of
healthy relatives was assembled from this group.
Followup visits. Each member of the cohort was inter-
viewed by telephone every 4 months, for 5 years. Each interview
was structured with an approximate duration of 7 minutes, and
divided into 4 components: 1) interviewer presentation, 2) the
COPCORD questionnaire, 3) a reminder for the subject to
telephone the research center immediately if experiencing joint
pain or if having doubts about having arthritis, and 4) answer-
ing any questions. The interviewers were senior medical stu-
dents who were trained by the principal investigator (CR-R)
and were blinded with regard to the IgM-RF and anti–CCP-2
results.
Subjects in the cohort were assessed in-office by partic-
ipating rheumatologists, who also were blinded with regard to
IgM-RF and anti–CCP-2 results, within 7 days in the event of a
“positive” result on the COPCORD questionnaire during the
telephone interview, or by direct request of the subject. The in-
office evaluation included administration of the COPCORD
questionnaire, joint assessment including a swollen joint count
and tender joint count based on 66 or 68 joints, and other pro-
cedures (e.g., laboratory or radiologic) deemed necessary
based on the rheumatologist’s judgment. A second appoint-
ment with the same rheumatologist was scheduled for within 2
weeks; if no data indicated the presence of any rheumatic dis-
ease, the patient remained in the cohort. If the rheumatologist
determined the presence of arthritis (swelling), participation
in the cohort was ended and the subject was followed up by the
rheumatologist until a specific diagnosis was obtained accord-
ing to the strategies proposed by van der Helm-van Mil et al
(28) and Yamane et al (29). Treatment decisions were made
by the rheumatologist.
Primary end point. The primary end point was defined
as fulfillment of the ACR 1987 criteria for the classification of
RA (26). However, all clinical charts were reviewed again in 2013
to ascertain RA classification according to the ACR/European
League Against Rheumatism (EULAR) 2010 criteria (13).
AUTOANTIBODY TESTING TO PREDICT RA DEVELOPMENT IN RELATIVES OF RA PATIENTS 2839

Table 1. Baseline demographic characteristics and frequency distribution of IgM-RF and anti-CCP levels by percentile in the 252 patients with
rheumatoid arthritis*
All Anti-CCP2/RF2 Anti-CCP2/RF1 Anti-CCP1/RF2 Anti-CCP1/RF1
(n 5 252) (n 5 39) (n 5 27) (n 5 15) (n 5 171) P
Age, mean 6 SD years 49 6 13 48 6 15 49 6 10 48 6 10 49.56 13 NS
Age at diagnosis, mean 6 SD years 39 6 13 43 6 16 36 6 10.5 40 6 10 38 6 13 NS
Female, no. (%) 223 (88) 35 (90) 23 (85) 15 (100) 150 (88) NS
Anti-CCP, units/ml
Mean 6 SD 91 6 107 0.3 6 1.2 0.9 6 2.2 57 6 60 130 6 109 ,0.01
25th percentile 5.2 0 0 24 38.6
50th percentile 44.3 0 0 41.5 102.2
75th percentile 200 0 0 53 200
RF, units/ml
Mean 6 SD 289 6 502 2.0 6 5.4 173 6 234 467 400 6 570 ,0.01
25th percentile 25 0 40 0 77.8
50th percentile 107 0 82 0 168.5
75th percentile 324 0 240 11.8 462

* IgM-RF 5 IgM rheumatiod factor; anti-CCP 5 anti2cyclic citrullinated peptide; NS 5 not significant.

Quality assurance. Source documents and databases
were revised by the principal investigator every 6 months. A
random sample of 10 telephone interviews was repeated by the
principal investigator every 4 months to assess the accuracy of
the data. Telephone interviews continued even if individuals
moved to another city; in these cases a local rheumatologist was
contacted and asked to follow study procedures if an in-office
evaluation was required.
This study was approved by 2 institutional review
boards (ethics committees): one from the hospital of the health
care system for workers in Mexico (Instituto Mexicano del
Seguro Social) and the other from the research center (Unidad
de Investigacion en Enfermedades Cronico-Degenerativas).
Statistical analysis. RA patients and family members
were categorized into 4 groups based on the results of anti–CCP-
2 and IgM-RF testing: 1) seronegative (anti–CCP-2–negative
and IgM-RF–negative), 2) anti–CCP-2–negative and IgM-RF–
positive, 3) anti–CCP-2–positive and IgM-RF–negative, and 4)
anti–CCP-2–positive and IgM-RF–positive. IgM-RF and anti–
CCP-2 were analyzed as nominal variables (positive or negative),
as continuous variables, and by percentiles.
The significance of differences in continuous variables
was determined by one-way analysis of variance or Kruskal-
Wallis test with Scheffe’s post hoc analysis for multiple com-
parisons. Pearson’s chi-square or Fisher’s exact test was used
for categorical variables. Ninety-five percent confidence inter-
vals (95% CIs) were calculated. P values less than or equal to
0.05 were considered significant.
Kaplan-Meier survival curves were used to assess RA
incidence. An event was defined as the time (months since
baseline) when subjects developed arthritis (swelling) of 1 or
more joints, and further fulfilled the ACR 1987 criteria. Cen-
sored data were defined if subjects continued to be asymptom-
atic or were lost to followup. A Cox analysis was used to assess
baseline risk factors predicting RA development over time.
Variables considered as risk predictors were 1) sociodemo-
graphic variables (age, sex, and consanguinity grade such as
parents, offspring, uncles/aunts, grandsons/granddaughters,
etc.), and 2) serologic markers (anti–CCP-2 and RF) at base-
line. A multivariable stepwise approach was used as follows: in
the first step, univariable analyses were performed for all vari-
ables; covariates that had a P value of ,0.1 or influenced the
regression coefficient of age/sex by .10% were selected for
step 2. All relevant covariates were analyzed together in step 3.
In order to assess confounding, in step 4 all significant variables
and confounders were analyzed. Three multivariable models
were computed according to anti–CCP-2 and IgM-RF status.
Statistical analyses were performed using Stata software.
RESULTS
Patients with RA. We invited 279 patients with
RA for this study, and 252 (90%) attended the appoint-
ment at the research center with at least 1 family member.
Table 1 shows demographic characteristics of the patients,
and frequency distribution of IgM-RF and anti–CCP-2
levels by percentile. Most of the RA patients were female,
and the mean disease duration was 10 years; 74% were
positive for anti–CCP-2, 79% for IgM-RF, and 68% for
both autoantibodies. Seventy patients (28%) brought one
relative, 64 (25%) brought 2 relatives, 83 (33%) brought
between 3 and 5 relatives, 30 (12%) between 6 and 10 rel-
atives, and 5 (2%) between 11 and 22 relatives. Five of the
initial 279 patients were from multicase RA families.
Family members in the cohort. RA patients
attended the research center with 825 first- and second-
degree biologic relatives; 6 were excluded (3 because blood
samples could not be drawn and 3 because they already
had a rheumatic disease, i.e., juvenile chronic arthritis,
CREST syndrome [calcinosis, Raynaud’s phenomenon,
esophageal dysmotility, sclerodactyly, telangiectasias], and
psoriatic arthritis in 1 each). Therefore, the cohort was
assembled with 819 healthy relatives. Most of the relatives
were female and were offspring of the RA patients. Eleven
(1.3%) were positive for both IgM-RF and anti–CCP-2.
The anti–CCP-2–positive/RF-negative group was an aver-
age of 8 years older compared with the seronegative
group and the anti–CCP-2–negative/RF-positive groups.
2840 RAMOS-REMUS ET AL

Table 2. Baseline characteristics of the 819 healthy relatives of patients with rheumatoid arthritis*
All Anti-CCP2/RF2 Anti-CCP2/RF1 Anti-CCP1/RF2 Anti-CCP1/RF1
(n 5 819) (n 5 780) (n 5 16) (n 5 12) (n 5 11) P
Age, mean 6 SD years 35 6 12 35 6 12 35 6 15 43 6 10.5 45 6 7.5 ,0.01
Female, no. (%) 562 (69) 528 (68) 15 (94) 10 (83) 9 (82) NS
Smoking, no. (%) 202 (25) 195 (25) 2 (13) 4 (33) 1 (9) NS
First-degree relatives, no. (%) 648 (79) 614 (79) 14 (88) 10 (83) 10 (91) NS
Parents 48 (6) 44 (6) 2 (13) 1 (8) 1 (9)
Siblings 265 (32) 248 (32) 4 (25) 5 (42) 8 (73)
Offspring 335 (41) 322 (41) 8 (50) 4 (33) 1 (9)
Second-degree relatives, no. (%) 171 (21) 166 (21) 2 (13) 2 (17) 1 (9) NS
Uncles/aunts 22 (3) 21 (3) 0 0 1 (9)
Nephews/nieces 91 (11) 89 (11) 1 (6) 1 (8) 0
Cousins 30 (4) 29 (4) 0 1 (8) 0
Grandchildren 28 (3) 27 (4) 1 (6) 0 0
Anti-CCP, units/ml
Mean 6 SD 3.2 6 23.2 0.1 6 0.4 0.2 6 1 86 6 76 141 6 88 ,0.01
25th percentile 0 0 0 17.4 26.9
50th percentile 0 0 0 69.1 200
75th percentile 0 0 0 155.1 200
RF, units/ml
Mean 6 SD 3.6 6 23 0.6 6 3 46 6 39 5.8 6 7.4 147.5 6 114 ,0.01
25th percentile 0 0 24.7 0 54.4
50th percentile 0 0 28.9 0 94.5
75th percentile 0 0 53.3 12.3 248

* See Table 1 for definitions.

First-degree relatives appeared to have a higher point Tables 3 and 4 and Figure 1 show outcomes dur-
prevalence of RA-related antibodies when compared with ing the 5-year period by serologic group. The best positive
second-degree relatives (Table 2). predictive value (PPV) occurred when both anti–CCP-2
The frequency distribution of the autoantibodies and IgM-RF were positive (63.6% [95% CI 35, 92]),
was assessed by age group. This analysis revealed posi- when compared to the groups positive for anti–CCP-2
tivity for IgM-RF and anti–CCP-2, respectively, in 1 only (58.3% [95% CI 30, 86]) or RF only (0%).
(1.4%) and 0 of 72 individuals younger than 21 years, in Seventeen subjects (2.1%) developed RA. The
6 (2.4%) and 1 (0.4%) of 254 individuals between 21 level of anti–CCP-2, when positive, did not appear to
and 30 years old, in 5 (2.7%) and 6 (3.3%) of 183 indi- affect its PPV (3 of the subjects who developed RA were
viduals between 31 and 40 years old, in 8 (4.3%) and anti-CCP negative; of the remaining 14, the level of anti-
8 (4.3%) of 187 individuals between 41 and 50 years old, CCP was in the 25th percentile in 4, the 50th percentile in
in 6 (5.7%) and 7 (6.7%) of 105 individuals between 51 2, and the 75th percentile in 8). Only 1 relative of RA
and 60 years old, and in 1 (5.6%) and 1 (5.6%) of 18 patients in the 5 multicase families developed the disease.
individuals older than 60 years. The median time to RA incidence by serologic
Five-year followup of the cohort. As per proto- group was 2.4 years and 4.5 years in the anti–CCP-2–
col, the last interview occurred in December 2012. During positive/RF-negative and anti–CCP-2–positive/RF-positive
the study period, 83 subjects (10%) were lost to followup groups, respectively. Only 0.4% and 0% of the family
(80 could not be located and 3 had died [1 of lung cancer, members in the seronegative group and the anti–CCP-2–
1 of myocardial infarction, and 1 of unknown cause]). negative/RF-positive groups, respectively, developed RA.
Table 3. Sensitivity, specificity, and predictive value by serologic group*
Anti-CCP2/RF2 Anti-CCP2/RF1 Anti-CCP1/RF2 Anti-CCP1/RF1
(n 5 780) (n 5 16) (n 5 12) (n 5 11)
Events, no. (%) 3 (0.4) 0 7 (58) 7 (64)
Incidence, per 1,000 subjects per year 0.9 0 205 262.5
Time at risk, person-years 3,312.6 67.7 34.2 26.7
Sensitivity, % (95% CI) 17.6 (20.5, 36) 0 50 (24, 76) 50.5 (24, 76)
Specificity, % (95% CI) 3.1 (1.9, 4.3) 98 (97, 99) 99 (99, 100) 99.5 (99, 100)
PPV, % (95% CI) 0.4 (0.0, 0.8) 0 58.3 (30, 86) 63.6 (35, 92)
NPV, % (95% CI) 64 (49, 79) 98 (97, 99) 99 (98.5, 100) 99 (98.5, 100)

* 95% CI 5 95% confidence interval; PPV 5 positive predictive value; NPV 5 negative predictive value (see Table 1 for other definitions).
AUTOANTIBODY TESTING TO PREDICT RA DEVELOPMENT IN RELATIVES OF RA PATIENTS 2841

Table 4. Sensitivity, specificity, and predictive value by anti-CCP and RF status*

Anti-CCP1 Anti CCP2 RF1 RF2
(n 5 23) (n 5 796) (n 5 27) (n 5 792)
Events, no. (%) 14 (61) 3 (0.4) 7 (26) 10 (1.3)
Incidence per 1,000 subjects per year 230 0.9 74.3 3.0
Time at risk, person-years 60.8 3,380 94.2 3,346.7
Sensitivity, % (95% CI) 82 (64, 100) 17.6 (5, 44) 50 (24, 76) 58.8 (33, 81)
Specificity, % (95% CI) 99 (98, 100) 1.1 (0.6, 2) 97.5 (96, 99) 2.5 (2, 4)
PPV, % (95% CI) 60.9 (41, 81) 0.4 (0.1, 1) 25.9 (9, 42.5) 1.3 (0.6, 3)
NPV, % (95% CI) 99.6 (99, 100) 39 (20.5, 61) 99 (98.5, 100) 74 (53, 88)

* 95% CI 5 95% confidence interval; PPV 5 positive predictive value; NPV 5 negative predictive value (see Table 1 for other
definitions).

Kaplan-Meier analysis revealed an overall annual
incidence of 4.9 cases of RA for every 1,000 relatives
studied, with 3,441 person-years of time at risk. The
incidence increased to 262.5 per 1,000 person-years
when both autoantibodies (anti–CCP-2 and RF) were
positive. The annual incidence when anti–CCP-2 was
positive, independent of RF, was 230 new cases per
1,000 person-years. In contrast, the annual incidence
when RF was positive, independent of anti–CCP-2, was
74.3 new cases per 1,000 person-years.
Table 5 shows results of the Cox analysis. RF and
anti–CCP-2 increased the risk of developing RA, and
the absence of both serologic markers was protective
against RA (hazard ratio 0.01 [95% CI 0.0, 0.03]). Off-
spring had 3 times more risk of developing RA over 5
years than other family members (hazard ratio 3.1 [95%
CI 1.2, 8.1]). Smoking had no significance in the univari-
ate analysis. Age, degree of relationship, and presence
of the studied autoantibodies in the related RA patients
were not significant in the multivariate models.
All 17 of the subjects who developed RA accord-
ing to the ACR 1987 criteria (26) also met the ACR/
EULAR 2010 criteria (13) based on chart review per-
formed during 2013. No other subject in the cohort
developed clinical signs of arthritis (i.e., swollen joint)
or other inflammatory rheumatic disease.
DISCUSSION
This 5-year followup cohort study was designed
to investigate the premise that it is possible to identify
healthy individuals who are at increased risk of develop-
ing RA, using simple laboratory tests commonly avail-
able in clinical settings, so the results could be obtained
during a regular medical consultation. We found that
the highest PPV for RA development was 63.6%, when

Figure 1. Five-year survival curve by serologic group. aCCP 5 anti–cyclic citrullinated peptide; RF 5 rheumatoid factor.
2842 RAMOS-REMUS ET AL

Table 5. Cox analysis of the incidence of rheumatoid arthritis*

Univariable analysis Multivariable analysis
Model 1: Model 2: Model 3:
22LL goodness- 22LL goodness- 22LL goodness-
of-fit 288.5, of-fit 276.4, of-fit 264.18,
HR (95% CI) P HR (95% CI) HR (95% CI) HR (95% CI)
Age (years) 1.0 (1.0, 1.1) ,0.01 NS † NS
Sex (female) 2.1 (0.6, 7.5) NS NA NA NA
Smoking (yes vs. no) 0.9 (0.3, 2.5) NS NA NA NA
Oral hormone users (yes vs. no) 1.7 (0.8, 3.6) NS NA NA NA
Relative status (offspring vs. others) 3.1 (1.2, 8.1) ,0.05 ‡ ‡ ‡
Serologic group (anti-CCP1,RF1 vs. others) 70.1 (26.5, 185.7) ,0.01 52.5 (19.1, 144.1) – –
Anti-CCP level (units/ml) 1.03 (1.02, 1.03) ,0.01 – 1.02 (1.01, 1.03) –
RF level (units/ml) 1.02 (1.01, 1.02) ,0.01 – 1.00 (1.00, 1.01) –
Anti-CCP positive ($7.0 units/ml) vs. negative 223.1 (63.8, 779.9) ,0.01 – – 207.3 (54.8, 784.8)
RF positive ($20 units/ml) vs. negative 23.1 (8.8, 60.7) ,0.01 – – †

* HR 5 hazard ratio; 95% CI 5 95% confidence interval; LL 5 log-likelihood ratio; NA 5 not applicable (because of nonsignificant P value in
univariate analysis) (see Table 1 for other definitions).
† Lost significance in step 3.
‡ Lost significance in step 2.

both IgM-RF and anti–CCP-2 were positive. However,
the performance of IgM-RF positivity in the absence of
anti–CCP-2 was poor. Offspring of RA patients also had
an independent risk of developing the disease.
Our study had several strengths. These included
the stringent definition of healthy status at baseline, the
use of the second-generation anti-CCP test (anti-CCP2,
which performs better than anti-CCP1) (30,31), the fol-
lowup process, the case ascertainment procedures, and
the low percentage of subjects lost to followup (10%).
However, there were potential limitations as well.
First, generalizability might be limited because our results
were from a Mexican Mestizo population, which may
have different genetic characteristics when compared
with other ethnic groups. Also, we decided not to assess
HLA–DRB1, HLA–DR3, or shared epitopes (32–34), so
as to allocate our limited resources to generating data
that can be obtained in routine clinical settings. Second,
although consanguinity (descent from the same ancestor)
with an RA patient was an inclusion criterion, we did not
restrict the sample by level of relationship. In addition,
we could not include entire families, only those able to
and interested in participating. This self-selection could
act as a bidirectional participation bias and explain the
broad range of the 95% confidence intervals of the PPVs
and the relatively small number of subjects who devel-
oped RA. Third, we used anti–CCP-2 as one of the out-
come predictors. It has already been established that
production of anti–CCP is strongly dependent on the
genetic background of the patient, with anti–CCP positiv-
ity associated with the presence of HLA–DRB1 (31). In
our study, the majority of relatives who developed RA
(14 of 17) were anti–CCP-2 positive. Thus, it appears that
the PPV of anti-CCP2 is for the development of RA asso-
ciated with HLA2DRB1 (in which there is generally a
higher rate of joint destruction) (28,35), but not with
HLA2DR3 (in which the frequency of joint erosions has
been reported to be lower) (36). However, we did not ana-
lyze subjects in the present study by HLA type, and our
results therefore do not enable direct confirmation of this.
It has been reported in several publications, from as
early as the late 1950s, that RF, and more recently anti-
CCP, is present years before RA develops (37,38). How-
ever, we found only 5 published cohort studies to contrast
with our results. In 1963, Ball and Lawrence (37) identified
subjects who initially did not have RA and observed them
for 5 years. Of the 19 subjects who were positive for RF at
high titers, 7 (36.8%) developed RA, while 4 (14.3%) of 28
subjects with no or low-titer RF developed the disease. In
a 1969 study, Mikkelsen and Dodge did not find any new
cases of RA among 105 asymptomatic subjects with RF
positivity who were observed for 3–6 years (39). However,
missing data on subject selection and followup procedures
in the two above-mentioned studies preclude direct com-
parison with ours. The largest population-based cohort
study was reported by del Puente et al in 1988 (40). They
conducted a 19-year followup of 2,712 Pima Indians in
Arizona who did not have RA at the time of enrollment.
Seventy new cases of RA developed; rates increased from
2.4 to 48.3 per 1,000 person-years, with the highest rates
among those who had the highest RF titers at baseline.
The single previously reported cohort study in
relatives of RA patients was published in 1992 by Silman
et al (41). They studied 370 previously unaffected first-
degree relatives from 65 multicase RA families in the
UK, with followup via letters sent annually to the subject
AUTOANTIBODY TESTING TO PREDICT RA DEVELOPMENT IN RELATIVES OF RA PATIENTS
and his or her general practitioner. The mean length of
followup was 5 years. Fourteen of these individuals
developed RA, but only 5 were positive for IgM-RF by
ELISA (34.8 per 1,000 person-years). Using the data
from that study, we calculated that the PPV for RF posi-
tivity was 21% (95% CI 9, 40); this is quite similar to
our result of 26% (95% CI 9, 43) (Table 4), despite the
fact that the RA families from which Silman and col-
leagues recruited relatives were all multicase RA fami-
lies and were of different ethnicity than our study
population. A report published in 2009 also described a
cohort of first-degree relatives of probands with RA
(25); however, only the baseline, cross-sectional analysis
of the data is reported (1.7% had anti-CCP and 4.7%
had IgM-RF). There are 5 other published reports of
studies on relatives of patients with RA (42–46), but the
cross-sectional design precludes direct comparison with
ours. Anti-CCP was found in between 5% and 8.5% of
the relatives in these studies, RF in between 14% and
33%, and both autoantibodies in 3.1% of first-degree
relatives.
There are also blood bank–based studies that
have demonstrated the presence of RF and/or anti-CCP
years prior to the diagnosis of RA, in US populations
(15,18,19) and in populations from Finland (14), Ice-
land (16), Sweden (21,22,47), The Netherlands (20),
and Norway (17). Additionally a study of patients with
undifferentiated arthritis demonstrated the utility of
anti-CCP status in predicting subsequent development
of RA (48).
Thus, there are only two reported cohort studies
in relatives of patients with RA: one in multicase RA
families, and the one we are reporting here. Further
investigations are needed to determine the PPV of RA-
related autoantibodies in other settings, but our present
data could already be used to design interventional stud-
ies in healthy family members of RA patients. For
instance, our results could contribute to screening strate-
gies to identify individuals at higher risk of RA develop-
ment, who might benefit from counseling interventions
to improve self-efficacy and coping strategies, to encour-
age avoidance of potential triggers of the disease such as
smoking and obesity, and to decrease time to referral in
order to begin treatment during the window of opportu-
nity. Indeed, it may also be appropriate to evaluate the
feasibility of clinical trials for primary prevention in high-
risk healthy relatives.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors approved
the final version to be published. Dr. Ramos-Remus had full access to
2843
all of the data in the study and takes responsibility for the integrity of
the data and the accuracy of the data analysis.
Study conception and design. Ramos-Remus, Castillo-Ortiz, Aguilar-
Lozano, Padilla-Ibarra, Sandoval-Castro, Vargas-Serafin, de la Mora-
Molina, Ramos-Gomez, Sanchez-Ortiz, Avila-Armengol, Aceves-Avila.
Acquisition of data. Ramos-Remus, Castillo-Ortiz, Aguilar-Lozano,
Padilla-Ibarra, Sandoval-Castro, Vargas-Serafin, de la Mora-Molina,
Ramos-Gomez, Sanchez-Ortiz, Avila-Armengol, Aceves-Avila.
Analysis and interpretation of data. Ramos-Remus, Castillo-Ortiz,
Aguilar-Lozano, Padilla-Ibarra, Sandoval-Castro, Vargas-Serafin, de
la Mora-Molina, Ramos-Gomez, Sanchez-Ortiz, Avila-Armengol,
Aceves-Avila.
REFERENCES
1. Badley EM. Rheumatic diseases: the unnoticed elephant in the
room. J Rheumatol 2008;35:6–7.
2. Klareskog L, Catrina AI, Paget S. Rheumatoid arthritis. Lancet
2009;373:659–72.
3. Loza E, Jover JA, Rodriguez-Rodriguez L, Carmona L.
Observed and expected frequency of comorbid chronic diseases
in rheumatic patients. Ann Rheum Dis 2008;67:418–21.
4. Ramos-Remus C, Sierra-Jimenez G, Skeith K, Aceves-Avila FJ,
Russell AS, Offer R, et al. Latitude gradient influences the age
of onset in rheumatoid arthritis patients. Clin Rheumatol 2007;
26:1725–8.
5. Mottonen T, Hannonen P, Korpela M, Nissila M, Kautiainen H,
Ilonen J, et al. for the FIN-RACo Trial Group. Delay to institu-
tion of therapy and induction of remission using single-drug or
combination–disease-modifying antirheumatic drug therapy in
early rheumatoid arthritis. Arthritis Rheum 2002;46:894–8.
6. Lard LR, Visser H, Speyer I, vander Horst-Bruinsma IE,
Zwinderman AH, Breedveld FC, et al. Early versus delayed
treatment in patients with recent-onset rheumatoid arthritis:
comparison of two cohorts who received different treatment
strategies. Am J Med 2001;111:446–51.
7. Bejarano V, Quinn M, Conaghan PG, Reece R, Keenan AM,
Walker D, et al. and the Yorkshire Early Arthritis Register Con-
sortium. Effect of the early use of the anti–tumor necrosis factor
adalimumab on the prevention of job loss in patients with early
rheumatoid arthritis. Arthritis Rheum 2008;59:1467–74.
8. Combe B, Landewe R, Lukas C, Bolosiu HD, Breedveld F,
Dougados M, et al. EULAR recommendations for the manage-
ment of early arthritis: report of a task force of the European
Standing Committee for International Clinical Studies Including
Therapeutics (ESCISIT). Ann Rheum Dis 2007;66:34–45.
9. Quinn MA, Emery P. Window of opportunity in early rheuma-
toid arthritis: possibility of altering the disease process with early
intervention. Clin Exp Rheumatol 2003;21 Suppl 31:S154–7.
10. Nell VP, Machold KP, Eberl G, Stamm TA, Uffmann M,
Smolen JS. Benefit of very early referral and very early therapy
with disease-modifying anti-rheumatic drugs in patients with ear-
ly rheumatoid arthritis. Rheumatology (Oxford) 2004;43:906–14.
11. Bykerk V, Emery P. Delay in receiving rheumatology care leads
to long-term harm [editorial]. Arthritis Rheum 2010;62:3519–21.
12. Van der Linden MP, le Cessie S, Raza K, van der Woude D, Kne-
vel R, Huizinga TW, et al. Long-term impact of delay in assessment
of patients with early arthritis. Arthritis Rheum 2010;62:3537–46.
13. Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT,
Bingham CO III, et al. 2010 rheumatoid arthritis classification
criteria: an American College of Rheumatology/European
League Against Rheumatism collaborative initiative. Arthritis
Rheum 2010;62:2569–81.
14. Aho K, Palosuo T, Raunio V, Puska P, Aromaa A, Salonen JT.
When does rheumatoid disease start? Arthritis Rheum 1985;28:
485–9.
15. Bos WH, Nielen MM, Dijkmans BA, van Schaardenburg D.
Duration of pre-rheumatoid arthritis anti-cyclic citrullinated pep-
2844
tide positivity is positively associated with age at seroconversion
[letter]. Ann Rheum Dis 2008;67:1642.
16. Jonsson T, Thorsteinsson J, Kolbeinsson A, Jonasdottir E,
Sigfusson N, Valdimarsson H. Population study of the impor-
tance of rheumatoid factor isotypes in adults. Ann Rheum Dis
1992;51:863–8.
17. Jorgensen KT, Wiik A, Pedersen M, Hedegaard CJ, Vestergaard
BF, Gislefoss RE, et al. Cytokines, autoantibodies and viral anti-
bodies in premorbid and postdiagnostic sera from patients with
rheumatoid arthritis: case-control study nested in a cohort of
Norwegian blood donors. Ann Rheum Dis 2008;67:860–6.
18. Karlson EW, Chibnik LB, Tworoger SS, Lee IM, Buring JE,
Shadick NA, et al. Biomarkers of inflammation and development
of rheumatoid arthritis in women from two prospective cohort
studies. Arthritis Rheum 2009;60:641–52.
19. Majka DS, Deane KD, Parrish LA, Lazar AA, Baron AE, Walker
CW, et al. Duration of preclinical rheumatoid arthritis-related
autoantibody positivity increases in subjects with older age at time
of disease diagnosis. Ann Rheum Dis 2008;67:801–7.
20. Nielen MM, van Schaardenburg D, Reesink HW, van de Stadt
RJ, van der Horst-Bruinsma IE, de Koning MH, et al. Specific
autoantibodies precede the symptoms of rheumatoid arthritis: a
study of serial measurements in blood donors. Arthritis Rheum
2004;50:380–6.
21. Rantapaa-Dahlqvist S, Boman K, Tarkowski A, Hallmans G. Up
regulation of monocyte chemoattractant protein-1 expression in
anti-citrulline antibody and immunoglobulin M rheumatoid fac-
tor positive subjects precedes onset of inflammatory response
and development of overt rheumatoid arthritis. Ann Rheum Dis
2007;66:121–3.
22. Rantapaa-Dahlqvist S, de Jong BA, Berglin E, Hallmans G,
Wadell G, Stenlund H, et al. Antibodies against cyclic citrullin-
ated peptide and IgA rheumatoid factor predict the development
of rheumatoid arthritis. Arthritis Rheum 2003;48:2741–9.
23. MacGregor AJ, Snieder H, Rigby AS, Koskenvuo M, Kaprio J,
Aho K, et al. Characterizing the quantitative genetic contribu-
tion to rheumatoid arthritis using data from twins. Arthritis
Rheum 2000;43:30–7.
24. Seldin MF, Amos CI, Ward R, Gregersen PK. The genetics rev-
olution and the assault on rheumatoid arthritis [review]. Arthritis
Rheum 1999;42:1071–9.
25. Kolfenbach JR, Deane KD, Derber LA, O’Donnell C, Weisman
MH, Buckner JH, et al. A prospective approach to investigating
the natural history of preclinical rheumatoid arthritis (RA) using
first-degree relatives of probands with RA. Arthritis Rheum
2009;61:1735–42.
26. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, et al. The American Rheumatism Association 1987
revised criteria for the classification of rheumatoid arthritis.
Arthritis Rheum 1988;31:315–24.
27. Cardiel MH, Rojas-Serrano J. Community based study to esti-
mate prevalence, burden of illness and help seeking behavior in
rheumatic diseases in Mexico City: a COPCORD study. Clin
Exp Rheumatol 2002;20:617–24.
28. Van der Helm-van Mil AH, le Cessie S, van Dongen H, Breedveld
FC, Toes RE, Huizinga TW. A prediction rule for disease out-
come in patients with recent-onset undifferentiated arthritis: how
to guide individual treatment decisions. Arthritis Rheum 2007;56:
433–40.
29. Yamane T, Hashiramoto A, Tanaka Y, Tsumiyama K, Miura Y,
Shiozawa K, et al. Easy and accurate diagnosis of rheumatoid arth-
ritis using anti-cyclic citrullinated peptide 2 antibody, swollen joint
count, and C-reactive protein/rheumatoid factor. J Rheumatol
2008;35:414–20.
30. Fernandez-Suarez A, Reneses S, Wichmann I, Criado R, Nunez
A. Efficacy of three ELISA measurements of anti-cyclic citrullin-
ated peptide antibodies in the early diagnosis of rheumatoid
arthritis. Clin Chem Lab Med 2005;43:1234–9.
RAMOS-REMUS ET AL
31. Van Venrooij WJ, van Beers JJ, Pruijn GJ. Anti-CCP antibody,
a marker for the early detection of rheumatoid arthritis. Ann
N Y Acad Sci 2008;1143:268–85.
32. De Vries RR, Huizinga TW, Toes RE. Redefining the HLA and
RA association: to be or not to be anti-CCP positive.
J Autoimmun 2005;25 Suppl:21–5.
33. Irigoyen P, Lee AT, Wener MH, Li W, Kern M, Batliwalla F,
et al. Regulation of anti–cyclic citrullinated peptide antibodies in
rheumatoid arthritis: contrasting effects of HLA–DR3 and the
shared epitope alleles. Arthritis Rheum 2005;52:3813–8.
34. Verpoort KN, van Gaalen FA, van der Helm-van Mil AH,
Schreuder GM, Breedveld FC, Huizinga TW, et al. Association
of HLA–DR3 with anti–cyclic citrullinated peptide antibody–
negative rheumatoid arthritis. Arthritis Rheum 2005;52:3058–62.
35. Van Oosterhout M, Bajema I, Levarht EW, Toes RE, Huizinga
TW, van Laar JM. Differences in synovial tissue infiltrates
between anti–cyclic citrullinated peptide–positive rheumatoid
arthritis and anti–cyclic citrullinated peptide–negative rheuma-
toid arthritis. Arthritis Rheum 2008;58:53–60.
36. Visser H, le Cessie S, Vos K, Breedveld FC, Hazes JM. How to
diagnose rheumatoid arthritis early: a prediction model for per-
sistent (erosive) arthritis. Arthritis Rheum 2002;46:357–65.
37. Ball J, Lawrence JS. The relationship of rheumatoid serum fac-
tor to rheumatoid arthritis: a 5-year follow up of a population
sample. Ann Rheum Dis 1963;22:311–8.
38. Lawrence JS, Ball J. Genetic studies on rheumatoid arthritis.
Ann Rheum Dis 1958;17:160–8.
39. Mikkelsen WM, Dodge H. A four year follow-up of suspected
rheumatoid arthritis: the Tecumseh, Michigan, Community
Health Study. Arthritis Rheum 1969;12:87–91.
40. Del Puente A, Knowler WC, Pettitt DJ, Bennett PH. The inci-
dence of rheumatoid arthritis is predicted by rheumatoid factor
titer in a longitudinal population study. Arthritis Rheum 1988;
31:1239–44.
41. Silman AJ, Hennessy E, Ollier B. Incidence of rheumatoid arth-
ritis in a genetically predisposed population. Br J Rheumatol
1992;31:365–8.
42. Smolik I, Robinson DB, Bernstein CN, El-Gabalawy HS. First-
degree relatives of patients with rheumatoid arthritis exhibit high
prevalence of joint symptoms. J Rheumatol 2013;40:818–24.
43. Kim SK, Bae J, Lee H, Kim JH, Park SH, Choe JY. Greater prev-
alence of seropositivity for anti-cyclic citrullinated peptide antibody
in unaffected first-degree relatives in multicase rheumatoid
arthritis-affected families. Korean J Intern Med 2013;28:45–53.
44. Young KA, Deane KD, Derber LA, Hughes-Austin JM, Wagner
CA, Sokolove J, et al. Relatives without rheumatoid arthritis show
reactivity to anti–citrullinated protein/peptide antibodies that are
associated with arthritis-related traits: studies of the etiology of
rheumatoid arthritis. Arthritis Rheum 2013;65:1995–2004.
45. Barra L, Scinocca M, Saunders S, Bhayana R, Rohekar S,
Racape M, et al. Anti–citrullinated protein antibodies in unaf-
fected first-degree relatives of rheumatoid arthritis patients.
Arthritis Rheum 2013;65:1439–47.
46. Arlestig l, Mullazehi M, Kokkonen H, Rocklov J, Ronnelid J,
Dahlqvist SR. Antibodies against cyclic citrullinated peptides of
IgG, IgA and IgM isotype and rheumatoid factor of IgM and
IgA isotype are increased in unaffected members of multicase
rheumatoid arthritis famiies from northern Sweden. Ann Rheum
Dis 2012;71:825–9.
47. Berglin E, Padyukov L, Sundin U, Hallmans G, Stenlund H, van
Venrooij WJ, et al. A combination of autoantibodies to cyclic
citrullinated peptide (CCP) and HLA-DRB1 locus antigens is
strongly associated with future onset of rheumatoid arthritis.
Arthritis Res Ther 2004;6:R303–8.
48. Van Gaalen FA, Linn-Rasker SP, van Venrooij WJ, de Jong
BA, Breedveld FC, Verweij CL, et al. Autoantibodies to cyclic
citrullinated peptides predict progression to rheumatoid arthritis
in patients with undifferentiated arthritis: a prospective cohort
study. Arthritis Rheum 2004;50:709–15.