The Chronic Fatigue Syndrome (CFS) disease classification attempts to identify patients with a condition distinct from the fatigue of other chronic illness. It is most frequently defined by the Fukuda definition: severe, unexplained, chronic fatigue lasting at least 6 months, accompanied by 4 of 8 symptom criteria (post-exertional malaise, impaired memory or concentration, unrefreshing sleep, muscle pain, arthralgia, headaches, sore throat or, tender cervical or axillary lymph nodes)[1, 2]. According to population based studies, 800,000 or more people in the US meet these criteria [3–8] or have CFS like disease [8]. The Center for Disease Control (CDC) posts that the illness costs the US $9 billion annually in lost earnings and productivity [9, 10].

CFS is a complex illness that is not well understood. This applies to the genetic and environmental origins of CFS, as well as factors that may contribute to the severity of symptoms and outcomes. Despite efforts to develop standardized research criteria to define the disease, progress in CFS diagnosis and treatment has been slow, in part due to concurrence in origins, causation, lack of a standard clinical definition, and lack of specific biomarkers. A community-based screening study by the CDC indicated that only 9%-16% of individuals with CFS had been diagnosed. Variations in presentation (age of onset, sudden versus gradual onset, and co-morbid conditions) further increase the heterogeneity of CFS patients. It has been proposed that predisposition to CFS may have a genetic basis, although environmental factors, viral illnesses, stressful life events or traumas may also be implicated [4, 6, 8, 11, 12]. CFS has been described as a disorder of the brain, the immune system, or the endocrine system; studies have documented underlying pathophysiological processes involving all of these systems [13–18].

There are numerous studies suggesting that genes might play a role in the development of CFS or in its progression [19–21]. Several twin studies have examined genetic vs. environmental determinants of CFS [21–24]. These studies have shown higher concordance in monozygotic vs. dizygotic twins. A small familial aggregation study (CFS cases = 25), showed excess relative risks for CFS in first-degree relatives of CFS cases compared to controls [20].

Analysis of a heritable contribution to CFS is confounded by the newness of the diagnosis, the lack of specific diagnostic criteria or biomarkers, and high rates of co-morbidity with other syndromes. Here we report the existence of a unique, population-based resource with genealogical information linked to diagnosis data allowing analysis of the genetic relationships among 811 individuals diagnosed with CFS since 1993. The combined data have been analyzed to examine the hypothesis of a heritable contribution to CFS.

The original Utah Population Database (UPDB) represented genealogical records of Utah pioneers and their descendants. The UPDB was created from family group sheets compiled by members of the Church of Jesus Christ of Latter-day Saints (LDS) submitted to the Genealogical Society of Utah. Researchers at the University of Utah computerized these genealogical paper records of Mormon pioneers and their descendants into an electronic database [25].

The original UPDB genealogy represented approximately1.6 million individuals in genealogies up to 6 generations deep. Genealogical data added since the original derives from state records (birth, death, driver's license and others) entailing 6.5 million individuals. New individuals from these sources who link to the original genealogy data produce pedigrees up to 15 generations deep with 2.3 million individuals belonging to genealogies at least 3 generations deep.

Numerous Utah data sets have been linked to UPDB records, including Utah Cancer Registry (UCR) records. The UPDB has also been linked to individual medical data from the University of Utah Health Sciences Hospital (UUHSH) Enterprise Data Resource Center. The UUHSH Data Warehouse (DW) contains comprehensive hospital, medical, and clinical data from 1993 to present. The UUHSH serves 20% of the state's residents. The UUHSH Data Resource Center currently contains over 1.8 million patient demographic records for patients treated at the hospital and outpatient clinics. Over 1.26 million (73%) of these hospital records have been matched to a "person record" in UPDB.

CFS patients were identified by the presence in the medical record of a CFS-specific ICD-9 diagnostic code (780.71) originating from medical providers involved in the patient's care. Although in other areas of the country the rate of CFS diagnosis may be lower, it is likely that in Utah the appropriate application of the Fukuda definition and diagnosis of CFS increases the rate. This is likely related to the CFS referral clinic in the community, and the availability of CFS-specific Continuing Medical Education to community providers during the last decade. UUHSH has no clinics specifically devoted to CFS, or specialists who claim expertise in CFS.

Although all University of Utah hospital and clinic records have been linked to the UPDB genealogy, only a selected subset of these records can be accessed, due to the sensitive nature of the data. A set of randomly selected hospital patients was identified for use as controls, representing 20% of all University of Utah hospital/clinic patients who also had linked genealogy data (n~200,000). We created cohorts to which these UUHSC patients were assigned, and from which controls were selected randomly. All individuals were assigned to 1 of 132 cohorts based on sex, birth-year, and birthplace (i.e., Utah or not). All controls for the analyses performed were selected randomly from this set.

This unique resource allowed us to identify all genetic relationships between patients with a diagnosis of CFS. To examine evidence for a genetic contribution to CFS in the UPDB, two different analyses were performed: estimation of relative risks in close and distant relatives of CFS cases, and comparison of average relatedness in CFS cases to controls.

We identified 941 patients with a CFS diagnosis in the UUHSC DW from 1993 to present that also had Utah genealogy data. Age at diagnosis was estimated using the minimum admission date for a CFS diagnosis, and thus may be over estimated. Twelve percent of the cases had an age at diagnosis less than age 20 years; 28% of the cases had an age at diagnosis greater than 59 years; 653 (69%) were female. Because CFS is associated with various co-morbid conditions, with possible genetic associations, we excluded the 130 CFS cases that also had a diagnosis of cancer in the UCR. This left 811 CFS cases for analysis.

Genealogical Index of Familiality

Table 2 shows the results of the GIF test for excess relatedness in the 811 CFS cases, including the average relatedness of the cases, the mean average relatedness for 1000 sets of matched controls, the empirical significance of the overall test for excess relatedness, and the empirical significance of the dGIF. The average relatedness of CFS cases was significantly greater than expected when all relationships were considered (p <0.001), showing strong evidence for excess familial clustering (due to either shared environment, or shared genes, or a combination). When close relationships (genetic distance <4) were ignored, the significant excess was still present (p = 0.010). Figure 1 displays the GIF test graphically. It shows the contribution to the GIF statistic (y-axis) by the relationship (x-axis; genetic distance) between all pairs for CFS cases (and controls). This observation of significant excess relationships observed at almost all degrees of relationship strongly supports a genetic contribution to predisposition to CFS.

Phenotype	n	Case Average Relatedness	Mean Control Average Relatedness	Empirical p-value GIF	Empirical p-value Distant GIF
CFS	811	3.15	2.31	< 0.001	0.010

The existence of a Utah resource combining up to 15 generations of genealogy data with medical diagnosis data from 1993 has allowed testing of the hypothesis of a heritable contribution to CFS. The methods used in this study have previously provided evidence for a heritable component to many diseases, including: prostate cancer, influenza mortality, aneurysm, cancer, rotator cuff disease, asthma mortality, and diabetes, among others [28, 32–34]. Studies of Utah high-risk pedigrees identified in the UPDB have lead to the discovery of multiple cancer predisposition genes including BRCA1, BRCA2, p16, and HPC2/ELAC2 [35–38]. The UPDB data analyzed represents a homogeneous population that has been shown to be genetically representative of Northern Europe, with normal U.S. inbreeding levels [39, 40].

Significantly increased risks among first degree relatives are often referred to as providing evidence for a "genetic" contribution to disease. However, given the sharing among close relatives of their genes, lifestyle, and environment, increased first degree risk may simply indicate familial clustering, it does not provide evidence for a genetic contribution. However, significant excess risks in second and third degree relatives strongly indicates a genetic contribution to disease, given the much lower likelihood of these relatives sharing common risks and environments.

Analysis of CFS in a large Utah resource shows clear evidence of significant excess familial clustering and significantly elevated risks for CFS among first, second, and third degree relatives of CFS cases. The results strongly support a genetic contribution to predisposition to CFS as it is currently defined and diagnosed by clinicians in Utah. Although a genetic predisposition to CFS has been suggested in the literature, this is the first population-based analysis to comprehensively support this claim.

This study used a uniform, consistent source for all diagnoses, and is not limited by bias introduced by study designs involving selected ascertainment of cases or requiring recall for diagnoses. The most significant limitation of this analysis is the narrow window of view to identify individuals diagnosed with CFS. This results from the relatively short period of time for which this diagnosis has existed, and the limited time-period of diagnosis data available (1993-present). These effects limit our ability to identify cases who might be related across different generations (e.g. grandparent/grandchild or avunculars). Although CFS cases may have been censored from our observation in this resource, cases are uniformly censored across the resource, leading to conservative, but unbiased, estimates of familiality.

We excluded CFS cases with a cancer diagnosis, which might have been a cause of CFS symptoms in these cases. Other potential confounders could not be considered: including other heritable predisposing conditions (e.g., depression), or risk factors that are familial, but not genetic (e.g., occupation, socioeconomic factors, or healthcare access).

This study of CFS heritability does not allow determination of the mechanisms that lead to predisposition to CFS. We have identified multiple pedigrees with a significant excess of CFS cases. We propose study of these pedigrees to identify the gene(s) predisposing to CFS, as well as to better understand mechanisms and potential environmental factors and triggers.

Studies reporting an association of CFS with XMRV and MLV-related viruses have provided conflicting results [41–44]. While association of CFS with an infectious-like syndrome at onset is recognized, and many microbial and viral infections have been implicated as possible triggers, no single agent has been associated with a large fraction of cases. It might be hypothesized that a heritable predisposition to virus infection explains both our findings and the complex virus associations that have been recognized.

Identification of CFS predisposition genes, and increased understanding of how these genes affect health could allow identification of predisposed individuals at an earlier age, prophylactic screening for at-risk individuals, improved healthcare standards to reduce risk of CFS development, all leading to identification of treatments or medications that could prevent or delay onset of symptoms in those impaired by this debilitating disease.