-
PDF
- Split View
-
Views
-
Cite
Cite
David M. Koelle, Serge Barcy, Meei-Li Huang, Rhoda L. Ashley, Lawrence Corey, Judy Zeh, Suzanne Ashton, Dedra Buchwald, Markers of Viral Infection in Monozygotic Twins Discordant for Chronic Fatigue Syndrome, Clinical Infectious Diseases, Volume 35, Issue 5, 1 September 2002, Pages 518–525, https://doi.org/10.1086/341774
Close -
Share
Abstract
To estimate the prevalence of viruses associated with chronic fatigue syndrome (CFS) and to control for genetic and environmental factors, we conducted a co-twin control study of 22 monozygotic twin pairs, of which one twin met criteria for CFS and the other twin was healthy. Levels of antibodies to human herpesvirus (HHV)-8, cytomegalovirus, herpes simplex virus 1 and 2, and hepatitis C virus were measured. Polymerase chain reaction (PCR) assays for viral DNA were performed on peripheral blood mononuclear cell specimens to detect infection with HHV-6, HHV-7, HHV-8, cytomegalovirus, Epstein-Barr virus, herpes simplex virus, varicella zoster virus, JC virus, BK virus, and parvovirus B19. To detect lytic infection, plasma was tested by PCR for HHV-6, HHV-8, cytomegalovirus, and Epstein-Barr virus DNA, and saliva was examined for HHV-8 DNA. For all assays, results did not differ between the group of twins with CFS and the healthy twins.
Chronic fatigue syndrome (CFS) is characterized by fatigue that lasts ⩾6 months and is accompanied by disturbances of sleep, cognitive function, musculoskeletal pain, and other symptoms [1]. It is a symptom-based diagnosis of exclusion without distinguishing findings for physical examination or routine laboratory tests [1, 2] that is reported most frequently in female subjects [3–5]. Frequent clinical features of CFS, such as abrupt onset with fever, adenopathy [6], and flulike symptoms [2], in combination with epidemiologic features consistent with outbreaks [7, 8], have suggested that CFS may be the result of a viral infection. Although previous studies have shown that various measurements of viral infection do not consistently distinguish patients with CFS from control subjects [9–17], these studies generally have not directly measured virus persistence or replication by use of nucleic acid detection. Rather, the focus has been on assessing seroprevalence or indirect indicators of viral replication, such as the presence of IgM antibodies. Furthermore, past research has not adequately controlled for the role of genetic or host factors and environmental exposures, making it difficult to draw definitive conclusions regarding the role of viruses in CFS.
One alternative to traditional study designs that compare CFS patients to healthy, depressed, or chronic disease groups of control subjects who are genetically dissimilar and have different exposure histories is a co-twin control study. This is a matched-pair design that adjusts perfectly for genetic factors and partially for environmental ones (which are typically not considered in case-control studies) [18]. We undertook a co-twin control study to examine the relation between CFS and markers of viral infection in 22 pairs of monozygotic twins who were discordant for CFS. We performed serologic assays for antibodies to human herpesvirus (HHV)-8, cytomegalovirus (CMV), herpes simplex virus (HSV) 1 and 2, and hepatitis C virus (HCV). We performed PCR assays for viral DNA that tested peripheral blood mononuclear cell (PBMC) samples to detect infection with viruses known to be latent in this compartment, including HHV-6, HHV-7, HHV-8, CMV, and Epstein-Barr virus (EBV), as well as viruses without established tropism for circulating PBMC, such as HSV, varicella zoster virus (VZV), JC virus, BK virus, and parvovirus B19. To detect lytic infection, plasma samples were tested by PCR for HHV-6, HHV-8, CMV, and EBV DNA, and saliva samples were examined for HHV-8 DNA.
Patients, Materials, and Methods
Registry construction and subject recruitment
Twin patients, one or both of whom reported persistent fatigue, were recruited for participation in a volunteer twin registry through patient support group newsletters (58% of participants), clinicians and researchers familiar with CFS (11%), notices placed on electronic bulletin boards (15%), twin organizations and researchers (6%), relatives and friends (3%), and other sources (8%). Intake questionnaires were mailed to 600 individuals, 426 (71%) of whom returned them. Complete data were available for both members of 193 twin pairs (386 individuals). The questionnaire collected information on demographics, zygosity, lifestyle, physical health conditions, the nature and extent of fatigue, and a checklist of CFS symptoms [1]. Twins who did not report persistent fatigue completed a control questionnaire that did not reference fatigue. All twins provided written informed consent. Details of the twin registry are available elsewhere [19].
Psychiatric disorders
The Diagnostic Interview Schedule Version III-A [20], a structured interview based on the Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised [21], was administered by telephone. Sections included major depression, dysthymia, generalized anxiety, panic, agoraphobia, posttraumatic stress disorder, mania, bipolar disorders, schizophrenia, eating disorders, somatization, and substance abuse and dependence.
Selection of the clinical sample
From the twin registry, 22 sets of monozygotic twins were chosen for a 6-day evaluation. Twins were required (1) to be ⩾18 years of age; (2) to have been reared together; (3) to be discordant for CFS (i.e., one twin met the Centers for Disease Control and Prevention criteria for CFS and the other was healthy and denied chronic fatigue); (4) to provide evidence of a recent negative HIV-1 antibody test; (5) to discontinue use of alcohol, caffeine, and all medications known to affect immune or inflammatory function ⩾2 weeks before the evaluation; and (6) to be able to travel to Seattle at the same time for evaluation.
All twins were rigorously screened. For twins with fatigue, the symptom checklist and diagnoses generated by the Diagnostic Interview Schedule were used to determine whether CFS criteria were met. Next, all medical records for the previous 5 years were reviewed by an internist for exclusionary conditions. Health issues were resolved by means of telephone or physician contact or laboratory tests before the evaluation. A psychologist and an infectious diseases specialist independently reviewed the charts and approved the twins for participation. Just before the scheduled visit, screening questions were read ministered to document that the ill twin still met CFS criteria and the other twin was healthy.
Zygosity was initially determined via validated [22, 23] self-report methods, then confirmed by restriction fragment-length polymorphism analysis. DNA was extracted from PBMC and digested with HaeIII, separated by agarose electrophoresis, blotted onto a nylon membrane, and hybridized with 6 variable-number tandem repeat probes to determine monozygosity with a certainty of ⩾0.999 [24].
Specimens, serological testing for viruses, and viral nucleic acid detection
Blood specimens were collected in anticoagulated tubes, for isolation of plasma and PBMC, and in non-anticoagulated tubes, for recovery of serum. PBMC (3 × 106 cells) were frozen in 200 µL of PBS. Plasma, PBMC, and saliva samples were stored at −80°C, and serum samples were stored at −20°C.
Antibodies to HHV-8 were detected as described elsewhere [25]. Total antibodies to CMV were detected by ELISA (Abbott). Antibodies to HSV-1 and HSV-2 were detected by immunoblot [26]. Antibodies to HCV were detected by ELISA (EIA 2.0, Abbott). IgM antibodies to parvovirus B19 were detected by ELISA (Specialty Labs).
Viral nucleic acid detection was performed on plasma, PBMC, and saliva samples. DNA was prepared from plasma and saliva (400 µL) with the Qiagen blood kit and from PBMC with the Qiagen tissue kit (Qiagen). DNA from saliva and blood was resuspended in 400 and 200 µL of water, respectively, with ∼10 µL of DNA per PCR reaction. For PBMC PCR, each reaction contained DNA extracted from 1.5 × 105 PBMC. For subjects with anti-HCV reactive serum, plasma RNA was prepared [27] and RNA detection performed by branched DNA testing (Quantiplex; Chiron).
DNA PCR was performed in duplicate. For all specimens, with the exception of those from the 2 most recently studied pairs, the PCR product was detected by means of liquid hybridization. Published methods and primers were used for HHV-6 [28], HHV-7 [29], HHV-8 [25], CMV [30], EBV [31], HSV [32], and BK virus [33]. For VZV, the primers were 5′-TCCGTTCTGGGTTCTGGTTGA-3′ and 5′-GCGCGGGGTCGCCTGATACTT-3′, and the probe was 5′-ACCACCCGGGCCCTGTGTTCG-3′. For JC virus, the primers were 5′-AGTCTTTAGGGTCTTCTACC-3′ and 5′-GGTGCCAACCTATGGAACAG-3′, and the probe was 5′-GATGATGAAAACACAGGATCCCAACACTC-3′. For parvovirus B19, the primers were 5′-TGAAAACTGGGCAATAAACTACAC-3′ and 5′-CTGCTGCTGATACTGGTGTCTGTG-3′, and the probe was 5′-TGCCCTCCACCCAGACCTCCAAACCA-3′. PCR inhibitors were detected with Drosophila controls, as described elsewhere [34]. Each PCR run included a standard control specimen containing plasmid DNA that included the target sequence in serial log10 dilutions, as well as a control containing no DNA. The lower limit of detection was generally 1–10 DNA copies per reaction. To be reported as having a “definite positive” result, both reactions needed to detect ⩾10 copies per reaction of target DNA. Specimens with <10 detectable copies and those for which only one of the duplicate PCRs was positive were reported as having a “low positive” result. If DNA was detected in the negative control, PCR was repeated.
For the most recent specimens, Taqman technology was used to detect PCR products, as described elsewhere, for HHV-6 [35], HHV-7 [29], HHV-8 [36], CMV [37], EBV [35], and HSV [38]. PCR methods for VZV and JC virus were similar. All BK virus assays used real-time PCR [33]. Inhibition of PCR was detected with an in-well coamplification control [36]. Briefly, each reaction well contained 50 copies of EXO DNA, 30 nM of primers EXO186F/315R, and 50 nM of probe EXO-P, which hybridized to the EXO PCR product and was labeled with VIC and TAMRA (6-carboxytetramethylrhodamine; all from Applied Biosystems). Wells that failed to yield a positive signal for EXO were considered to contain a PCR inhibitor; in this circumstance, the DNA was extracted a second time, and PCR was done again. Control specimens with 5000 template copies were processed with each specimen batch to ensure DNA recovery. To monitor for false positive results, 1× PBS was coprocessed from the DNA extraction stage through the PCR stage.
Statistical analysis
Initial descriptive analysis assessed mean differences by the matched paired t test. Because the number of twins discordant for virus detection was too small to permit the use of McNemar's test, exact binomial intervals for the probability that the CFS twin was positive for virus were used for the discordant pairs to obtain 95% CIs for ORs comparing the healthy and CFS twins [39]. If the 95% CI for the OR included 1, the CFS twins did not differ significantly from those without CFS. Analyses were done with S-Plus (Insightful) and SPSS (SPSS) software.
Results
Subject recruitment and selection
Of 193 twin pairs screened, 119 (62%) were discordant for ⩾6 months of fatigue. Of these 119 pairs, 67 (56%) were monozygotic and had complete data available; however, 14 pairs were not included because of exclusionary psychiatric illnesses, 4 because of medical disorders, and 1 because of a body mass index >45 [40]. In 9 pairs, the twin with fatigue did not meet CFS symptom criteria. In 4 pairs, the twin without fatigue had a condition exclusionary for CFS (e.g., cancer), and 6 pairs were excluded for other reasons (e.g., inadequate English-language skills and pregnancy). Of the remaining 29 pairs in which one twin met the criteria for CFS and the other twin was healthy and denied chronic fatigue, 22 (76%) completed the study, 1 (3%) refused to participate, 2 (7%) could not be scheduled, and 4 (14%) could not discontinue medications. There were no differences in physical or mental functional status, as measured by the Short Form-36 mental health scale [41], between the 22 twin pairs with CFS who traveled to Seattle for evaluation and the 7 who did not.
Demographic and clinical characteristics
The twins' mean age at the time of the study was 41.4 years; 20 pairs were women; and all twins were white and had been raised together. There was no difference between the group of twins with CFS and twins without CFS in the proportion that were married (59% vs. 59%). There was a small, statistically significant difference in the average years of schooling (twins with CFS, 14.0 years; twins without CFS, 14.7 years; P ⩽.01). The twins with CFS had a mean duration of illness of 7.0 years.
Virus assays
The most frequently detected infecting viruses overall were CMV and HSV-1 (20 [45%] of 44 for each) and HSV-2 (7 [16%] of 44) (table 1). HHV-8 infection was detected in 2 healthy twins. One had antibodies to latent, but not lytic, HHV-8 antigens. The other had antibodies to latent and lytic HHV-8 antigens and was shedding HHV-8 in saliva. One healthy twin with a reactive test for antibodies to HCV had normal transaminase levels, no history of hepatitis, and undetectable HCV RNA. ORs for comparison of serologic test results for healthy twins with those for twins with CFS did not differ significantly from 1.0.
Detection of viral infection in serum samples obtained from monozygotic twins who were discordant for chronic fatigue syndrome (CFS).
| Subject group | HHV-8 | CMV | HSV-1 | HSV-2 | HCV |
| By CFS status | |||||
| Twins with CFS (n = 22) | 0 | 8 | 12 | 4 | 0 |
| Healthy twins (n = 22) | 2a | 12 | 8 | 3 | 1b |
| All twins (n = 44) | 2 | 20 | 20 | 7 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) (n = 44) | 0 (0–3.5) | 0.3 (0–1.9) | 2.3 (0.5–14) | 1.3 (0.2–9.1) | 0 (0–19) |
| By viral assay result | |||||
| Negative for both twins (n = 44) | 20 | 8 | 7 | 15 | 21 |
| Positive for both twins (n = 44) | 0 | 6 | 5 | 0 | 0 |
| Positive for twin with CFS and negative for healthy twin (n = 44) | 0 | 2 | 7 | 4 | 0 |
| Positive for healthy twin and negative for twin with CFS (n = 44) | 2 | 6 | 3 | 3 | 1 |
| Subject group | HHV-8 | CMV | HSV-1 | HSV-2 | HCV |
| By CFS status | |||||
| Twins with CFS (n = 22) | 0 | 8 | 12 | 4 | 0 |
| Healthy twins (n = 22) | 2a | 12 | 8 | 3 | 1b |
| All twins (n = 44) | 2 | 20 | 20 | 7 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) (n = 44) | 0 (0–3.5) | 0.3 (0–1.9) | 2.3 (0.5–14) | 1.3 (0.2–9.1) | 0 (0–19) |
| By viral assay result | |||||
| Negative for both twins (n = 44) | 20 | 8 | 7 | 15 | 21 |
| Positive for both twins (n = 44) | 0 | 6 | 5 | 0 | 0 |
| Positive for twin with CFS and negative for healthy twin (n = 44) | 0 | 2 | 7 | 4 | 0 |
| Positive for healthy twin and negative for twin with CFS (n = 44) | 2 | 6 | 3 | 3 | 1 |
NOTE. Data are no. of subjects, unless indicated otherwise. CMV, cytomegalovirus; HCV, hepatitis C virus; HHV, human herpesvirus; HSV, herpes simplex virus.
In 1 twin, who shed HHV-8 DNA in saliva, latent and lytic antibodies were detected.
Plasma specimens were negative for HCV RNA.
Detection of viral infection in serum samples obtained from monozygotic twins who were discordant for chronic fatigue syndrome (CFS).
| Subject group | HHV-8 | CMV | HSV-1 | HSV-2 | HCV |
| By CFS status | |||||
| Twins with CFS (n = 22) | 0 | 8 | 12 | 4 | 0 |
| Healthy twins (n = 22) | 2a | 12 | 8 | 3 | 1b |
| All twins (n = 44) | 2 | 20 | 20 | 7 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) (n = 44) | 0 (0–3.5) | 0.3 (0–1.9) | 2.3 (0.5–14) | 1.3 (0.2–9.1) | 0 (0–19) |
| By viral assay result | |||||
| Negative for both twins (n = 44) | 20 | 8 | 7 | 15 | 21 |
| Positive for both twins (n = 44) | 0 | 6 | 5 | 0 | 0 |
| Positive for twin with CFS and negative for healthy twin (n = 44) | 0 | 2 | 7 | 4 | 0 |
| Positive for healthy twin and negative for twin with CFS (n = 44) | 2 | 6 | 3 | 3 | 1 |
| Subject group | HHV-8 | CMV | HSV-1 | HSV-2 | HCV |
| By CFS status | |||||
| Twins with CFS (n = 22) | 0 | 8 | 12 | 4 | 0 |
| Healthy twins (n = 22) | 2a | 12 | 8 | 3 | 1b |
| All twins (n = 44) | 2 | 20 | 20 | 7 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) (n = 44) | 0 (0–3.5) | 0.3 (0–1.9) | 2.3 (0.5–14) | 1.3 (0.2–9.1) | 0 (0–19) |
| By viral assay result | |||||
| Negative for both twins (n = 44) | 20 | 8 | 7 | 15 | 21 |
| Positive for both twins (n = 44) | 0 | 6 | 5 | 0 | 0 |
| Positive for twin with CFS and negative for healthy twin (n = 44) | 0 | 2 | 7 | 4 | 0 |
| Positive for healthy twin and negative for twin with CFS (n = 44) | 2 | 6 | 3 | 3 | 1 |
NOTE. Data are no. of subjects, unless indicated otherwise. CMV, cytomegalovirus; HCV, hepatitis C virus; HHV, human herpesvirus; HSV, herpes simplex virus.
In 1 twin, who shed HHV-8 DNA in saliva, latent and lytic antibodies were detected.
Plasma specimens were negative for HCV RNA.
There were no differences between the twins with CFS and the healthy twins with respect to the rate of detection of latent viruses (table 2). Overall, β-herpesviruses were the most common infecting virus detected in PBMC. HHV-7 was detected in 28 (64%) of 44 specimens; 86% of the positive results were classified “definite positive.” HHV-6 was detected in the PBMC of 14 (32%) of 44 twins (36% of twins with CFS; 27% of twins without CFS). CMV DNA was detected in 3 (7%) of 44 specimens, including 2 from a twin with CFS and 1 from a healthy twin. For 1 twin with CFS, CMV DNA was detected at 52 copies per reaction in 2 separate aliquots of PBMC, but no antibodies to CMV were detected. Latent γ-herpesvirus infections were infrequently detected. Overall, 9 twins, including 3 (14%) with CFS and 6 (27%) healthy twins, had EBV detected in PBMC. Of these, 1 twin with CFS and 2 healthy twins had definite positive results. Except for parvovirus B19 DNA in 1 healthy twin, no other virus DNA was detected in any specimen. For all viruses detected in ⩾1 twin, we calculated ORs for detection of latent virus in twins with CFS compared with healthy twins. ORs did not differ significantly from 1.0 (table 3).
Detection of viral DNA in samples of peripheral blood mononuclear cells obtained from monozygotic twins who were discordant for chronic fatigue syndrome (CFS).
| No. of patients, by virus | ||||||||||
| Subject group, viral assay result | HHV-6 | HHV-7 | HHV-8 | CMV | EBV | HSV | VZV | JC virus | BK virus | Parvovirus B19 |
| Twins with CFS (n = 22) | ||||||||||
| Any positive | 8 | 14 | 0 | 2 | 3 | 0 | 0 | 0 | 0 | 0 |
| Positive | 6 | 12 | 0 | 2 | 1 | 0 | 0 | 0 | 0 | 0 |
| Low positive | 2 | 2 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 |
| Healthy twins (n = 22) | ||||||||||
| Any positive | 6 | 14 | 0 | 1 | 6 | 0 | 0 | 0 | 0 | 1a |
| Positive | 3 | 12 | 0 | 1 | 2 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 3 | 2 | 0 | 0 | 4 | 0 | 0 | 0 | 0 | 0 |
| All twins (n = 44) | ||||||||||
| Any positive | 14 | 28 | 0 | 3 | 9 | 0 | 0 | 0 | 0 | 1 |
| Positive | 9 | 24 | 0 | 3 | 3 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 5 | 4 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 0 |
| No. of patients, by virus | ||||||||||
| Subject group, viral assay result | HHV-6 | HHV-7 | HHV-8 | CMV | EBV | HSV | VZV | JC virus | BK virus | Parvovirus B19 |
| Twins with CFS (n = 22) | ||||||||||
| Any positive | 8 | 14 | 0 | 2 | 3 | 0 | 0 | 0 | 0 | 0 |
| Positive | 6 | 12 | 0 | 2 | 1 | 0 | 0 | 0 | 0 | 0 |
| Low positive | 2 | 2 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 |
| Healthy twins (n = 22) | ||||||||||
| Any positive | 6 | 14 | 0 | 1 | 6 | 0 | 0 | 0 | 0 | 1a |
| Positive | 3 | 12 | 0 | 1 | 2 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 3 | 2 | 0 | 0 | 4 | 0 | 0 | 0 | 0 | 0 |
| All twins (n = 44) | ||||||||||
| Any positive | 14 | 28 | 0 | 3 | 9 | 0 | 0 | 0 | 0 | 1 |
| Positive | 9 | 24 | 0 | 3 | 3 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 5 | 4 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 0 |
NOTE. CMV, cytomegalovirus; EBV, Epstein-Barr virus; HCV, hepatitis C virus; HHV, human herpesvirus; HSV herpes simplex virus; VZV, varicella zoster virus.
Serologic tests were negative for IgM.
Detection of viral DNA in samples of peripheral blood mononuclear cells obtained from monozygotic twins who were discordant for chronic fatigue syndrome (CFS).
| No. of patients, by virus | ||||||||||
| Subject group, viral assay result | HHV-6 | HHV-7 | HHV-8 | CMV | EBV | HSV | VZV | JC virus | BK virus | Parvovirus B19 |
| Twins with CFS (n = 22) | ||||||||||
| Any positive | 8 | 14 | 0 | 2 | 3 | 0 | 0 | 0 | 0 | 0 |
| Positive | 6 | 12 | 0 | 2 | 1 | 0 | 0 | 0 | 0 | 0 |
| Low positive | 2 | 2 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 |
| Healthy twins (n = 22) | ||||||||||
| Any positive | 6 | 14 | 0 | 1 | 6 | 0 | 0 | 0 | 0 | 1a |
| Positive | 3 | 12 | 0 | 1 | 2 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 3 | 2 | 0 | 0 | 4 | 0 | 0 | 0 | 0 | 0 |
| All twins (n = 44) | ||||||||||
| Any positive | 14 | 28 | 0 | 3 | 9 | 0 | 0 | 0 | 0 | 1 |
| Positive | 9 | 24 | 0 | 3 | 3 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 5 | 4 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 0 |
| No. of patients, by virus | ||||||||||
| Subject group, viral assay result | HHV-6 | HHV-7 | HHV-8 | CMV | EBV | HSV | VZV | JC virus | BK virus | Parvovirus B19 |
| Twins with CFS (n = 22) | ||||||||||
| Any positive | 8 | 14 | 0 | 2 | 3 | 0 | 0 | 0 | 0 | 0 |
| Positive | 6 | 12 | 0 | 2 | 1 | 0 | 0 | 0 | 0 | 0 |
| Low positive | 2 | 2 | 0 | 0 | 2 | 0 | 0 | 0 | 0 | 0 |
| Healthy twins (n = 22) | ||||||||||
| Any positive | 6 | 14 | 0 | 1 | 6 | 0 | 0 | 0 | 0 | 1a |
| Positive | 3 | 12 | 0 | 1 | 2 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 3 | 2 | 0 | 0 | 4 | 0 | 0 | 0 | 0 | 0 |
| All twins (n = 44) | ||||||||||
| Any positive | 14 | 28 | 0 | 3 | 9 | 0 | 0 | 0 | 0 | 1 |
| Positive | 9 | 24 | 0 | 3 | 3 | 0 | 0 | 0 | 0 | 1 |
| Low positive | 5 | 4 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 0 |
NOTE. CMV, cytomegalovirus; EBV, Epstein-Barr virus; HCV, hepatitis C virus; HHV, human herpesvirus; HSV herpes simplex virus; VZV, varicella zoster virus.
Serologic tests were negative for IgM.
Assessments of twin pairs with virus detected in only 1 twin and of twins with chronic fatigue syndrome (CFS) versus healthy twins.
| Subject group, by viral assay result and/or CFS status | HHV-6 | HHV-7 | CMV | EBV | Parvovirus B19 |
| Positive for twin with CFS and negative for healthy twin | 4 | 4 | 1 | 2 | 0 |
| Negative for twin with CFS and positive for healthy twin | 2 | 4 | 0 | 5 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) | 2.0 (0.3–22) | 1.0 (0.2–5.4) | ∞ (0.05–∞) | 0.4 (0–2.4) | 0 (0–19) |
| Subject group, by viral assay result and/or CFS status | HHV-6 | HHV-7 | CMV | EBV | Parvovirus B19 |
| Positive for twin with CFS and negative for healthy twin | 4 | 4 | 1 | 2 | 0 |
| Negative for twin with CFS and positive for healthy twin | 2 | 4 | 0 | 5 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) | 2.0 (0.3–22) | 1.0 (0.2–5.4) | ∞ (0.05–∞) | 0.4 (0–2.4) | 0 (0–19) |
NOTE. Data are no. (%) of patients, unless indicated otherwise. CMV, cytomegalovirus; EBV, Epstein-Barr virus; HHV, human herpesvirus.
Assessments of twin pairs with virus detected in only 1 twin and of twins with chronic fatigue syndrome (CFS) versus healthy twins.
| Subject group, by viral assay result and/or CFS status | HHV-6 | HHV-7 | CMV | EBV | Parvovirus B19 |
| Positive for twin with CFS and negative for healthy twin | 4 | 4 | 1 | 2 | 0 |
| Negative for twin with CFS and positive for healthy twin | 2 | 4 | 0 | 5 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) | 2.0 (0.3–22) | 1.0 (0.2–5.4) | ∞ (0.05–∞) | 0.4 (0–2.4) | 0 (0–19) |
| Subject group, by viral assay result and/or CFS status | HHV-6 | HHV-7 | CMV | EBV | Parvovirus B19 |
| Positive for twin with CFS and negative for healthy twin | 4 | 4 | 1 | 2 | 0 |
| Negative for twin with CFS and positive for healthy twin | 2 | 4 | 0 | 5 | 1 |
| Comparison of twins with CFS vs. healthy twins, OR (95% CI) | 2.0 (0.3–22) | 1.0 (0.2–5.4) | ∞ (0.05–∞) | 0.4 (0–2.4) | 0 (0–19) |
NOTE. Data are no. (%) of patients, unless indicated otherwise. CMV, cytomegalovirus; EBV, Epstein-Barr virus; HHV, human herpesvirus.
Lytic herpesvirus replication was not commonly detected. Viral DNA sequences from HHV-6, HHV-8, CMV, or EBV were not detected in any plasma specimens, but 1 healthy twin had HHV-8 DNA detected in saliva specimens. This same subject had serum antibodies to both lytic and latent HHV-8 antigens.
Discussion
Most reports on the association between infection and CFS have examined infection with enteroviruses, retroviruses, or members of the Herpesviridae. Usually acquired early in life, herpesvirus infection remains latent throughout life, with periodic reactivation. The detection of lytic replication in adults, therefore, usually reflects reactivation rather than primary infection, and lytic replication may increase with decreased cellular immunity. In previous studies of EBV, increased levels of antibodies to early antigen or levels of IgM antibodies to viral capsid antigens have been noted and interpreted as indicating active viral replication from a primary infection or reactivation of latent virus [6, 42–46]. One study even demonstrated spontaneous B cell outgrowth (a surrogate method of EBV culture) for 30% of patients with CFS and 8% of control subjects [47]. However, studies of the clinical utility of measuring EBV antibody titers [48] and of the correlation of symptom severity or recovery with antibody titers [49] and a serologic survey of active infection in CFS outbreaks [50] have had findings inconsistent with a primary role for EBV in CFS. In this study, which focused on quantitative detection of viruses as indicators of active infection, we observed no differences between monozygotic twins discordant for CFS. However, the frequency that herpesvirus infection was detected by use of PBMC was likely an underestimate of the true biologic prevalence. EBV DNA was detected in only 20% of PBMC specimens, although most adults in the United States are latently infected with EBV [51]. The most likely reason for our low detection rate is that DNA from a relatively low number of B lymphocytes was present in the sample used in each PCR reaction [52, 53].
HHV-6 infection [54] also has been implicated in CFS, although its role remains controversial. Viral cultures have been positive for HHV-6 for some patients [15] but not all [16]. In addition, culture bioassays have demonstrated HHV-6 replication [7] and increased HHV-6 DNA levels [16]. As with EBV, significantly higher antibody titers to HHV-6 have been observed in patients with CFS than in control subjects [7, 14, 16, 55, 56], again with considerable overlap between the 2 groups. Nonetheless, although we, like others [57], occasionally detected HHV-6, our data do not support an association with CFS. Plasma samples were negative for HHV-6, a finding consistent with absence or very low levels of lytic HHV-6 replication.
HHV-7 has been isolated from persons with CFS and persons in good health [58, 59]. We frequently detected HHV-7 in PBMC, as have other investigators [60], but we found no association with CFS. We did not detect lytic HHV-7 infection. The finding that 2 (4.5%) of 44 twins were infected with HHV-8 is consistent with findings in the US population [36, 61, 62].
Last, CMV infection also has been considered a possible factor in CFS. Two studies reported that antibody titers to CMV were not significantly higher in patients with CFS than in control subjects [14, 55], but neither reported direct measurements of CMV replication. In our study, CMV was not detected in the plasma of any twin. However, 3 twins had detectable CMV in PBMC; one of these twins was seronegative. Because serologic tests have imperfect sensitivity and the findings of the plasma DNA PCR were negative, it is unlikely this twin had acute CMV infection.
Several other viruses were sought, but there was no evidence of infection with the papovaviruses, JC virus (associated with neurologic syndromes [63]), or BK virus (associated with renal disease in immunocompromised persons [33]) in PBMC. One twin had a PCR positive for parvovirus B19, an infective agent linked with chronic anemia [64]. This subject was not anemic and did not have IgM antibodies to parvovirus; thus, the significance of the PCR finding is unclear. Finally, 1 (2.3%) of 44 twins had antibodies to HCV, but plasma samples were negative for HCV RNA. This was not unexpected, because the seroprevalence of HCV in the United States is ∼1.8%, and up to 25% of infected persons can clear HCV RNA [65]. Other microorganisms implicated in CFS include Chlamydia and enterovirus [66, 67], but we did not obtain anatomically appropriate specimens to test for the presence of these infective agents.
Our decision to evaluate twins deserves comment. Because the results of assays to detect viral infection can be influenced by many factors, the selection of control subjects is critical. Little research has been conducted on viral infections or antiviral immune responses in twins. One study reported that antibody titers to EBV, CMV, and HSV were similar in twins [68]. Other studies have observed that genetic influences on the degree to which a host is permissive to viral infection [69, 70] and genetic influences on immune system responses to viruses are best controlled for by the study of monozygotic twins. In our study, each control subject was identical to the paired case patient with respect to age, sex, family environment, and background. Furthermore, the pairs of twins were free of medication and had blood samples obtained within minutes of each other, and assays were performed in batch fashion on the same day.
This co-twin control study also has several notable limitations. First, the method used to identify the patient sample was not ideal. Solicitation by advertisement resulted in a volunteer sample of twin pairs with the potential for ascertainment bias. However, the more desirable strategy of systematically identifying twins from a well-defined population-based twin registry is not readily accomplished in the United States. Thus, it is not known how representative the twins in this study were of either twins in general or of persons with CFS. Although highly selected in one respect, the twins with CFS were not recruited from tertiary-care referral centers, unlike most subjects in previous studies. Moreover, the demographic and clinical characteristics of our twins with CFS were similar to those of clinical populations described elsewhere [2, 4]. Nonetheless, patterns of viral infection could differ between subgroups of patients, such as those with briefer illnesses. A second limitation is related to sample size. Although we found little evidence that viral infection has a role in perpetuating CFS, clearly some patients develop the illness as a result of an infection. The characteristics of these patients remain to be defined by larger studies. Third, because the twins in our study were adults, primarily women, recruited from community medical practices, and because they met strict criteria for the presence of CFS, it may not be possible to generalize our results to other patient samples and more specialized clinical settings.
In summary, this co-twin control study of 22 monozygotic twin pairs discordant for CFS did not demonstrate a major contribution for viral infection in perpetuating CFS. This study does not exclude the possibility that infectious agents trigger the illness. Because a twin study adjusts for heritable and some environmental features and is well suited to the study of illnesses for which appropriate comparison groups are not clearly defined [18], it is an appropriate way to examine the strength of factors reportedly associated with CFS. Future research should examine the role of viruses in selected samples of patients, especially patients in the early stages of their fatigue illness.
Acknowledgments
We thank the participants in the University of Washington Twin Registry for their cooperation, patience, and goodwill and Dr. Leigh Sawyer, Program Officer, National Institute of Allergy and Infectious Diseases, for her encouragement and support. We also acknowledge our advisory panel, who, with sage advice and ongoing encouragement, improved our scientific efforts. Finally, we acknowledge Dr. Jack Goldberg for his careful review of the article in manuscript and his many helpful suggestions throughout this project.
References
Author notes
Financial support: National Institutes of Health (grant U19 AI38429 to D.B.).
- polymerase chain reaction
- herpesvirus 4, human
- simplexvirus
- chronic fatigue syndrome
- environmental factors
- hhv-6
- bk virus
- dna
- dna, viral
- erythema infectiosum
- human herpesvirus 1
- human herpesvirus 3
- herpesvirus 7, human
- human herpesvirus 8
- jc virus
- human parvovirus b19 virus
- plasma
- saliva
- twins
- monozygotic twins
- virus diseases
- infections
- antibodies
- cytomegalovirus
- genetics
- viruses
- hepatitis c virus
- peripheral blood mononuclear cell
- lytic phase

Comments