SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Objective

The etiology of Wegener's granulomatosis (WG) supposedly involves interplay between genetic susceptibility and environmental triggers. However, little is known about whether WG actually clusters in families. Information on the degree of familial aggregation in WG is of clinical relevance, because patients with WG often want to know whether their diagnosis puts their closest relatives at increased risk of the disease. The aim of this study was to investigate the risk of WG in relatives of patients with WG.

Methods

Using Swedish nationwide registers on morbidity, family structure, and vital status, we compared the occurrence of WG (register-based plus chart review) among 6,670 first-degree relatives and 428 spouses of 1,944 Swedish patients with WG with the occurrence among 68,994 first-degree relatives and 4,812 spouses of 19,655 control subjects from the general population. Relative risks were estimated using the Cox proportional hazards regression model.

Results

Two of the 6,670 first-degree relatives of patients with WG and 13 of the 68,994 first-degree relatives of their population controls had WG, resulting in a relative risk of 1.56 (95% confidence interval 0.35–6.90). None of the 428 spouses of patients had WG.

Conclusion

In absolute terms, the occurrence of WG among close biologic and nonbiologic relatives of patients with WG is low. In terms of relative risk, our results provide strong evidence against a pronounced increase in familial risk such as that noted for systemic lupus erythematosus, irritable bowel disease, and multiple sclerosis but are compatible with familial aggregation of a magnitude similar to that for rheumatoid arthritis.

For several chronic inflammatory diseases, the finding of a familial aggregation of cases has suggested the existence and importance of genetic factors. In rheumatoid arthritis (RA), a familial aggregation is well documented, candidate genes have been identified, and several environmental factors as well as gene–environment interactions have recently been discovered (1).

In analogy with RA, Wegener's granulomatosis (WG) is also supposed to result from an interplay between genetic susceptibility (2–5) and environmental triggers (6–9). However, considerably less is known about the degree of familial aggregation in WG. Our knowledge of the familial occurrence of WG is based entirely on case reports (10–14). As recently pointed out by other investigators (15), a better understanding of the possible inheritance of (susceptibility to) WG is imperative for our understanding of the impact of genetics on disease etiology. An assessment of the familial risk of WG is also important from a clinical point of view, because patients typically want to know whether their own diagnosis of WG translates to an increased risk of the disease in their closest relatives.

To assess the occurrence and relative risk (RR) of WG among close relatives of patients with the disease, we took advantage of the Swedish public and population-based health care system, data from the high-quality Swedish register on health and vital status, and the possibility to identify biologic and nonbiologic relatives through register linkage. We performed a cohort study of 6,670 biologic first-degree relatives and 428 spouses of 1,944 patients with WG, and 68,994 biologic first-degree relatives and 4,812 spouses of 19,655 control subjects from the general population.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Setting.

The public Swedish health care system, the national registers, and the national registration numbers (NRNs) allowing deterministic record linkage have all been described elsewhere (16). In this study, the following national, population-based, virtually complete registers were used. The Inpatient Register contains individual-based information on all inpatient care provided since 1964. For each individual, hospital discharge information on the diagnosis is recorded according to the International Classification of Diseases (ICD) versions 7, 8, 9, and 10 (17). The Register of Total Population contains official census data; all residents in Sweden who are alive at the end of each year are included, with NRNs, names, place of domicile, along with date of deaths for individuals who died during the year. Information on immigration and emigration is also included. The Swedish Multi-Generation Register provides information on vertical and horizontal first-degree relatives (parents, siblings, and children) for residents born in 1932 or thereafter. The Cause of Death Register includes the date of death, the main and contributory causes of death coded according to the ICD systems, and the NRNs for all residents deceased since 1952.

Subjects.

In the Inpatient Register, we identified all 2,288 patients discharged with a diagnosis of WG (ICD-8 code 446.2, ICD-9 code 446.4, or ICD-10 code M31.3) from 1970 to 2003. Fifty-three percent of these patients were male, the median year of entry into the cohort was 1995 (10th, 90th percentiles 1980–2003), and their mean age at the time of the first hospitalization listing a diagnosis of WG was 61 years.

General population controls.

For every patient with WG, we randomly selected 10 general population control subjects from the Population Register; the control subjects were matched for sex, year of birth, county of residence, and marital status (Figure 1).

thumbnail image

Figure 1. Procedure for register linkage used in the current study. 1 = identification of patients with Wegener's granulomatosis (WG); 2 = identification of their population controls; 3 = identification of first-degree relatives and spouses of patients and their controls; 4 = linkage to identify WG in, and vital status of, all subjects.

Download figure to PowerPoint

Biologic and nonbiologic first-degree relatives of patients and control subjects.

Using NRNs, we linked the patients with WG and their controls to the Multi-Generation Register. For 1,939 of the 2,288 patients with WG identified in the register, we identified at least 1 first-degree relative (1,586 parents, 1,274 siblings, and 3,810 offspring). For the 19,606 population controls, we identified 15,422 parents, 13,303 siblings, and 40,269 offspring. Taking birth cohort into account, the proportion of individuals for whom a parent could be identified was similar for patients, their controls, and for the register as a whole (18). Similarly, distribution of the number of identified siblings per index individual was similar between patients and their controls (data not shown), but the average number of offspring was slightly lower among patients (1.67 children per patient) than among control subjects (1.76 children per subject). Through linkage of the patients with WG and their controls to the Population Register, we also identified all spouses (defined through marriage) from 1975 to 2003. These individuals were included as nonbiologic, shared-environment controls (Figure 1).

Followup.

All patients, control subjects, first-degree relatives, and spouses were linked to the nationwide and complete Total Population Register and the Cause-of-Death Register. Among all 75,664 individuals, 1,324 (1.7%) were lost to followup or had to be excluded due to data ambiguities. There was no difference in the length of followup between the relatives of patients and the relatives of control subjects. On average, parents were followed up for 20 years, siblings were followed up for 28 years, offspring were followed up for 26 years, and spouses were followed up for 26 years. In total, the study encompassed 2,281,652 person-years of followup of relatives.

Familial WG.

To identify cases of WG among all first-degree relatives of patients and control subjects, we linked all of these relatives to the nationwide Inpatient Register, thereby identifying all individuals in these cohorts who had at least 1 hospital discharge with (but not necessarily primarily because of) a diagnosis of WG.

Chart review.

Previous validation of 78 individuals in the cohort of patients with WG revealed a diagnostic correctness of 87% (21). To reduce the risk of misclassification bias in the present study, one of us (AK) assessed the register-based diagnoses against information in the medical files. First, the diagnoses of WG in those relatives registered with this disease and in their respective index patients (none of the control subjects were registered with WG) were reviewed. Individuals not fulfilling the American College of Rheumatology and Chapel Hill Consensus Conference criteria for WG (22, 23) (no relatives of patients and 5 relatives of control subjects) were reclassified as not having WG. In the 5 relatives of control subjects who did not fulfill the criteria for WG, either WG had been misdiagnosed or another diagnosis had been erroneously coded as WG. Second, and in order to detect any additional cases of WG misdiagnosed as any other vasculitic disease in the register, the files of relatives registered with any systemic vasculitides were reviewed. No additional cases of WG were observed. Characteristics of the cohorts are presented in Table 1. The study was approved by the regional ethics review board in Stockholm.

Table 1. Characteristics of the Swedish patients with WG and their first-degree relatives and spouses, and the general population controls and their first-degree relatives and spouses*
 WG patientsPopulation controls
No.Male/female sex, %Birth year, median (IQR)No.Male/female sex, %Birth year, median (IQR)
  • *

    WG = Wegener's granulomatosis; IQR = interquartile range.

Patients1,94455/451931 (1921–1945)19,65555/451931 (1921–1944)
Parents1,58649/511917 (1909–1929)15,42248/521917 (1909–1930)
Siblings1,27451/491949 (1942–1958)13,30351/491948 (1942–1958)
Children3,81052/481959 (1949–1969)40,26952/481959 (1950–1970)
Spouses42835/651949 (1943–1956)4,81238/621948 (1943–1956)

Statistical analysis.

To assess the association between exposure (being related to a patient with WG rather than a control subject) and outcome (WG in the relative), we used a Cox proportional hazards regression model (19). In the analyses, subjects were considered at risk from the year of birth or from the year 1975, whichever came later, and up to censoring due to emigration, death, diagnosis of WG, or the year 2004, whichever came first. RRs were estimated for all relatives combined and separately for parents, siblings, offspring, and spouses. Subset analyses by attained age (younger or older than age 50 years) of the relative were also performed. Familial clustering and intracluster dependence were taken into account by using the robust sandwich estimator (20). The Cox regression models were estimated using SAS version 9.1 software (PROC PHREG) (SAS Institute, Cary, NC).

RESULTS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Following register linkage and chart review, we identified 2 cases of WG (1 pair of first-degree relatives [mother and son]) among the relatives of the patients with WG. Among the 68,994 relatives of the population controls, we identified 14 cases of WG (3 parents, 6 siblings, 4 offspring, 1 spouse; none of these cases occurred in the same family). The calculated incidence of WG in the relatives of the control subjects in this study was 0.66 per 100,000 (95% confidence interval [95% CI] 0.37–1.08). The corresponding RRs of WG among the first-degree relatives of patients with the disease were 1.56 (95% CI 0.35–6.90) for all biologic relatives combined, 3.36 (95% CI 0.35–32.0) for parents, 0 for siblings, 2.56 (95% CI 0.29–22.8) for offspring, and 0 for spouses (Table 2).

Table 2. Cases of WG among first-degree relatives of Swedish patients with WG and first-degree relatives of population controls*
 Relatives of patientsRelatives of controlsRR (95% CI)
  • *

    Values are the number of cases/number of relatives. WG = Wegener's granulomatosis; RR = relative risk; 95% CI = 95% confidence interval.

All first-degree relatives2/6,66813/68,9941.56 (0.35–6.90)
 Parents1/1,5863/15,4223.36 (0.35–32.0)
 Siblings0/1,2746/13,3030.0
 Children1/3,8104/40,2692.56 (0.29–22.8)
 Spouses0/4281/4,8120.0

DISCUSSION

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

In this large population-based study, we found that in absolute terms, the risk for WG is low. Indeed, among ∼7,000 first-degree relatives and >400 spouses of almost 2,000 patients with WG, we identified only 1 pair of first-degree relatives in which both relatives were affected with the disease, and no affected spouses. In terms of RRs, our results are compatible with a degree of familial increased risk, of a magnitude similar to that reported for RA (1.5 < RR < 5) (24–26) but essentially rule out a more pronounced familial risk (RR > 6) such as the RRs reported for several other autoimmune inflammatory diseases, including multiple sclerosis (MS), irritable bowel disease (IBD) (27, 28), psoriasis (29, 30), and systemic lupus erythematosus (SLE) (31, 32).

To date, there is little to suggest that susceptibility to WG rests on a single penetrant genetic marker (33). Instead, investigators in several studies have proposed or investigated candidate genes (15, 34–39), most of which are also common in other autoimmune diseases (40–43), suggesting a complex inheritance and room for gene–environment interactions.

Apart from genetic susceptibility, environmental factors may also lead to clustering of disease. With respect to WG, results of previous studies have suggested silica (8, 44, 45) as well as organic solvents and farming (8) as possible environmental risk factors. The absence of marked familial clustering of disease in our study would suggest that regardless of what environmental factors are involved in the etiology of WG, they seem to occur on an individual rather than a household level.

Some limitations of our study should be discussed. Despite the exceedingly large nationwide cohort of patients with WG accrued over many years, and although our data provide strong evidence against familial aggregation in the same order of magnitude as in MS, IBD, SLE, or psoriasis (27–28, 30, 46), the statistical precision around the observed 56% increase in risk was limited.

With respect to diagnostic correctness, we validated the medical files of all relatives and patient–relative pairs registered with WG but did not check the diagnostic correctness of all patients registered with the diagnosis of WG. Previous validations of 78 individuals in a cohort of patients with a diagnosis code of WG revealed a diagnostic correctness of 87%. Within the framework of the current study, we reviewed 36 files of patients registered with WG and observed a diagnostic correctness in the same order of magnitude. However, the 10–15% misclassification of WG leaves room for a potential underestimation of (relative) risks of the same magnitude. In our study, this would increase the RR from 1.56 to 1.75 and, hence, alters neither the clinical nor the etiologic conclusion.

The strengths of our study include the nationwide population-based setting that allowed us to identify a large cohort of patients with WG and a large set of control subjects from the general population and also allowed an unbiased identification of their relatives and spouses. Although the possibility of identifying relatives was dependent on the birth cohort, such that offspring could be identified only for individuals born before 1932, this criterion applied similarly to patients and their controls. A particular strength was our independent assessment of WG among relatives, which did not rely on self-report but on chart diagnoses reported prospectively to a national health register and which revealed an incidence well in accordance with previously reported incidences of WG (47–50). All register-based outcomes were further validated before the analyses were performed. Although in our study, the outcomes of patients and control subjects were identified on the basis of hospitalization (similar for patients and control subjects), the overwhelming proportion of patients in Sweden who had WG during the study period were hospitalized at least once, for diagnostic or therapeutic reasons. Indeed, in a previous assessment, all 74 patients with WG who were identified on the basis of antineutrophil cytoplasmic antibody testing had been treated as inpatients early after the diagnosis of WG (Knight A: unpublished observations).

In conclusion, our results suggest that the risk of WG among close relatives of patients with the disease is low and do not support extensive clustering of cases in families. At the same time, our results are compatible with some increased familial occurrence of the disease and, thereby, with a genetic susceptibility to the disease.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES

Dr. Knight had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Knight, Sandin, Askling.

Acquisition of data. Knight, Sandin, Askling.

Analysis and interpretation of data. Knight, Sandin.

Manuscript preparation. Knight, Sandin, Askling.

Statistical analysis. Sandin.

REFERENCES

  1. Top of page
  2. Abstract
  3. SUBJECTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. REFERENCES