http://www.retrovirology.com/content/7/1/57/comments
What patient group are they studying?
Robin Durham (23 July 2010) n/a
I am a patient. I was given the diagnosis 15 years ago using the original Fukuda definition (which does include tender lymphadenopathy despite what the authors of the study say). I also meet the criteria for the Canadian Consensus definition. I would have been excluded from this study because I have some of the physical findings of the Canadian definition. If I don't have CDC defined CFS, what do I have?
Competing interests
Patient
top
Methodological deficiencies in the study by Switzer et al
Gerwyn Morris (23 July 2010) University of Ulster
A number of people here have already commented (quite eloquently) on the inappropriateness of using the Reeves 2005 criteria to diagnose people with ME/CFS. Indeed, using a selection criteria which has roughly the same chance of separating people with this illness from a generally fatigued population as tossing a coin is not the most auspicious of starts. That said however there are also some serious shortcomings with the methodology of the Switzer et al. paper which would make the detection of an in vivo virus highly unlikely, if present, in the people included in this study, regardless of the patient criteria used.
The terms analytical sensitivity and diagnostic sensitivity, although often used interchangeably, have entirely different meanings. Analytical sensitivity establishes the lowest theoretical level of detection of the assay target. This may be determined empirically via serial dilutions.
When such an assay is applied to a population for isolation and or diagnostic purposes however, another kind of sensitivity becomes the dominant concern (Saah et al., 1997). Diagnostic sensitivity refers to the ability to detect a target substance in a processed sample from a person known to be infected.
The diagnostic sensitivity of the Switzer team's assays were not established. This despite the fact that the CDC was sent blood samples from people known to be infected by XMRV. That they chose not to use these samples to establish the diagnostic sensitivity of their assays seems counterproductive, that is, if the goal was to successfully detect the virus. This criticism applies to both their PCR assays and their Western Blot approach. Finally, the decision to send samples to the Koch institute whose own ELISA approach has not established diagnostic sensitivity served no scientific purpose.
Although important methodological problems with the Switzer et al. study are highlighted above, the following is perhaps the most significant:
Dr. Norbert Bannert, one of the co-authors of the Switzer et al. study, was also involved in a 2000 study entitled "Transspecies Transmission of the Endogenous Koala Retrovirus" (Fiebig et al., 2000). This paper was concerned with the detection of an endogenous gammaretrovirus in koalas that induces leukemias and immune deficiencies associated with opportunistic infections, such as chlamydiosis. Due to obvious similarities, its methodology should therefore have offered some relevant lessons for those attempting to detect XMRV in patient samples.
For the Fiebig et al. study a new KoRV isolate was obtained from mitogen-stimulated peripheral blood mononuclear cells (PBMCs). This was a technique used in the Lombardi et al. (2009) study but not by the CDC, despite Dr. Bannert's previous experiences in successfully isolating a gammaretrovirus using this technique.
Furthermore, to study the host range of KoRV, human 293 kidney cells and the human T lymphocyte lines C8166 and CEM, as well as rat and mouse fibroblasts (rat1 and NIH 3T3, respectively), were used. Provirus integration was shown by PCR. This was the same technique that was used for dramatically increasing proviral concentration in the Lombardi et al. study before PCR, and once again not used in this CDC study. One must ask why.
Finally, in the Fiebig et al. study the entire env gene (p15E and gp70) was amplified from the DNA of
the animal from which KoRVD-B was isolated and sequenced, and one antiserum was generated by immunizing with the recombinant ectodomain of p15E of KoRV. Therefore in the KoR study the antisera was produced from an infected host. The CDC had the opportunity to use such a method rather than rely on a synthetic clone but, mysteriously, declined to use the infected blood samples they were
sent by the Whittemore Peterson Institute.
Note that in the Fiebig et. al. KoRV study:
Wistar rats were inoculated with cell-free KoRV (grown either on rat1 cells or 293 cells) or with KoRV-producing rat1 cells. Eleven of 12 animals were positive for p15E-specific antibodies, and four animals showed high levels of provirus integration in PBMCs at day 21 [...] indicating a productive infection in all 11 animals. The cell-associated virus load decreased, however, and 63 days postinoculation, no provirus was detected in the PBMCs of all inoculated rats.
Despite this, coincubation of cell-free plasma and mitogen-stimulated PBMCs from these animals with 293 cells yielded infectious virus [...]. When organs (spleen, ovary, lymph node, lung, liver, and kidney) or PBMCs from two rats [...] were analyzed for provirus integration on day 70, no KoRV sequences were detected. (Fiebig et al., 2000)
Thus over two months post-infection, the proviral copy number in their PBMC as well as in organs known to be viral reservoirs had fallen too low to be detected by PCR; yet infectious virus was still isolated when their PBMC were activated and cultured.
The implication here is that, in their XMRV study, it seems that Switzer et al. went ahead with a PCR approach despite the fact that at least one of the study's authors knew about the extreme improbability of detecting proviral DNA in a host chronically infected by a gammaretrovirus. The technique used in the Fiebig et al. study was developed to isolate PERV (Tache et al, 2000). Considering that the CDC have engaged in considerable study of PERV because of the potential dangers to transplant patients and the blood supply from this virus, their non-adherence to proven methodology for isolating a gammaretrovirus seems cavalier. Afterall, the potential levels of infection of healthy controls by XMRV demonstrated in peer reviewed studies indicate a potential medical catastrophe in the making.
Many studies report failures to detect proviral DNA in hosts known to be infected. False negatives frequently occur when primers are not properly complementary, frequently due to intraspecies genetic variability. Klein et al. (1999) comment that a major mismatch of three or four nucleotide bases causes a complete failure of PCR.
Loussert-Ajaka and others reported a lack of sensitivity in detecting HIV2 provirus because of low proviral load. In fact "PCR negative" is a diagnostic category describing HIV infections in which the proviral titre is too low to be detected by PCR (Woodfall et al, 1992). Therefore, given the above information, expecting PCR to detect a low titre gammaretrovirus without determining the diagnostic sensitivity of the assay would appear to be at the very least unrealistic.
In similar vein, two of the reasons for the failure of serological assays are genetic heterogeneity and antigenic variability (Kaplan, 2003). This heterogeneity and variability is a particular problem in assaying MuLV class viruses. In a 1980 study by O'Donnell and Nowinsky, a panel of monoclonal antibodies specific to the Gp70 and p15E epitopes of the env protein in ecotropic MuLV was used in an attempt to investigate the serological properties of a dulatropic MuLV of the same strain. No antibody reactions occurred. Upon investigation this was determined to have been caused by very minor changes in the Gp70 and P15E epitopes (O'Donnell and Nowinsky, 1980).
Given this information, the failure of the CDC to even to attempt to establish the diagnostic sensitivity of their serological tests when they had the means to do so is a major departure from the scientific method and makes the results obtained quite meaningless. In fact the decision to send the samples to
another continent for one of the serology tests maximized the chances of a false negative result because of the possibility of different strains existing in geographically distant parts of the world.
As a final comment, the recent criticism of the Switzer et al. paper made by Dr. Susan Vernon deserve mention, as they highlight further methodological difficulties with this study:
The samples from these three study cohorts were collected using different types of tubes, each of which has a distinct way of being processed. [N]one of the blood tubes used were of the same type used in the Lombardi study. (They used tubes containing sodium heparin that are intended for use with virus isolation). The blood tubes from the 18 Georgia registry patients are designed to collect whole blood and preserve nucleic acid; it is not clear where the plasma came from for these subjects since plasma cannot be obtained using these blood tube types. So the explanation for not finding XMRV in these samples is simple this was a study designed to not detect XMRV using a hodge-podge sample set. (Vernon, 2010)
References:
Fiebig U, Hartmann MG, Bannert N, Kurth R, and Denner J. Transspecies Transmission of the Endogenous Koala Retrovirus. J Virol. 2006 June; 80(11): 56515654.
Kaplan M, Reasons for False Negative (Seronegative) Test Results in Lyme Disease. 2003.
http://www.anapsid.org/lyme/lymeseroneg.html
Klein D, Janda P, Steinborn R, Mller M, Salmons B, Gnzburg WH. Proviral load determination of different feline immunodeficiency virus isolates using real-time polymerase chain reaction: influence of mismatches on quantification. Electrophoresis. 1999 Feb; 20(2): 291-9.
Lombardi VC, Ruscetti FW, Das Gupta J, Pfost MA, Hagen KS, Peterson DL, Ruscetti SK, Bagni RK, Petrow-Sadowski C, Gold B, Dean M, Silverman RH, Mikovits JA. Detection of an infectious retrovirus, XMRV, in blood cells of patients with chronic fatigue syndrome. Science. 2009 Oct 23; 326(5952): 585-9. Epub 2009 Oct 8.
Loussert-Ajaka, Simon F, Farfara I, Houhou N, Couto-Fernandez J, Dazza MC and Brun-Vzinet F. Comparative study of single and nested PCR for the detection of proviral HIV2 DNA. INSERM. 1994 May 10.
O'Donnell PV, and Nowinski RC. Serological analysis of antigenic determinants on the env gene oroducts of AKR dualtropic (MCF) murine leukemia viruses. Virology. 1980 Nov; 107(1):81-88.
Saah, Alfred,J. and Hoover Donald R,Annals of Internal Medicine:"Sensitivity" and "Specificity Reconsidered:The meanings of these terms in Analytical and Diagnostic settings. 1997 Jan 1; 126(1): 91-94.
Switzer WM, Jia H, Hohn O, Zheng H, Tang S, Shankar A, Bannert N, Simmons G, Hendry RM, Falkenberg VR, Reeves WC, Heneine W. Absence of evidence of Xenotropic Murine Leukemia Virus-related virus infection in persons with Chronic Fatigue Syndrome and healthy controls in the United States. Retrovirology. 2010 Jul 1; 7(1): 57.
Tacke SJ, Kurth R, Denner J. Porcine endogenous retroviruses inhibit human immune cell function: risk for xenotransplantation? Virology. 2000 Mar 1; 268(1): 87-93.
Vernon S. Blood from a Stone. 2010 Jul 1. http://www.cfids.org/xmrv/070110study.asp
Woodfall B, Schechter MT, Le TN, Craib KJ, Cassol S, Montaner JS, O'Shaughnessy MV. Low viral load as defined by negative PCR is associated with slower progression of HIV disease. Int Conf AIDS. 1992 Jul 19-24; 8.
Competing interests
no competing interests