Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Is rheumatoid arthritis genetic? No question.
- Alex J MacGregor, Jerry Lanchbury, Alan S. Rigby, Jaakko Kaprio, and Harold Snieder (2 March 2002)
-
Not really a criticism
- Charles M. Grossman (23 May 2002)
-
There is a difference between bias and expectation
- Anders Jørgen Svendsen, Niels V. Holm, Kirsten Kyvik, Per Hyltoft Petersen and Peter Junker (5 June 2002)
-
rheumatoid arthritis: stress-induced.
- dr.manan vasenwala (24 July 2003)
|
|
|||
|
Alex J MacGregor, Arthritis Research Campaign Senior Fellow Twin Research and Genetic Epidemiology Unit, St Thomas' Hospital, London SE1 7EH, Jerry Lanchbury, Alan S. Rigby, Jaakko Kaprio, and Harold Snieder
Send response to journal:
|
The cover of BMJ 324 (7332) poses the question: ‘Rheumatoid arthritis: is it genetic?’ and answers ‘probably not’. We feel that this is neither a valid response nor indeed is it a valid question. The Danish twin study on which this assessment is based was a retrospective study of 37,338 twins from two birth cohorts. A total of 13 monozygotic (MZ) and 36 dizygotic (DZ) twins were identified in which either twin had rheumatoid arthritis (RA) (1). None of the MZ twin pairs was concordant for the disease while only 2 DZ pairs were concordant. The low proportion of concordant pairs and the lack of significance in difference between MZ and DZ concordance is taken as indicating that environmental influences predominate in explaining the occurrence of the disease. It is informative to consider the power on which this conclusion is based more explicitly. Applying standard modelling approaches to the data given for confirmed cases of RA in the Danish sample, we calculate that the study had 80% power to detect a heritability of 65% and 90% power to detect a heritability of 75%. (setting alpha=0.05 in a model containing additive genetic and unique environmental effects). Thus, the study lacked sufficient power to detect a genetic influence on RA equivalent to the 60% heritability estimated in the two most recent twin studies of the disease (2). It is suggested that the study’s design ensured that the point estimates of concordance were likely to be unbiased in spite of the lack of power. We are less convinced about this. The prevalence of confirmed RA in the Danish sample at 0.15% appears low in comparison with contemporary published estimates of RA prevalence in European populations (3), even after accounting for the numbers of younger twins. This points to potential deficiencies in the screening methods used to identify probands. The absence of the middle-aged birth cohort severely limits the representativeness of the sample. The failure to identify MZ concordant pairs in the study is in itself a problem. MZ twin pairs concordant for RA exist (4;5), hence the concordance estimate of zero reported here is, by definition, biased. The problems of undertaking studies to estimate heritability for rare diseases have been recognised for decades and inspired the designs of the two largest and most recent twin studies of RA conducted in Finland in 1986 (6) and the UK in 1993 (7). In their critical review of these two studies, Svendsen et al (1) incorrectly highlight a number of potential methodological limitations that require further comment. Contrary to their assertion, cases of ankylosing spondylitis were specifically excluded in the Finnish study (and have been published (8)). While the Finnish register for medications has limitations, the diagnoses have been shown to be reliable. The attending physician, who makes the application, is usually a specialist and there are specified criteria for each disease that have to be fulfilled before reimbursement for medication costs is possible. The UK study used a parallel recruitment strategy involving the identification of probands in a defined time interval from both GP clinics and through a media campaign. Although a relatively higher proportion of MZ compared with DZ probands was ascertained, there was no reason to suspect that MZ concordant pairs were over-represented. The similarity of concordance rates among the GP and media recruited samples supports this conclusion. Both the Finnish and UK studies yielded significant estimates of heritability for RA of approximately 60% (2). The similar heritability estimates derived from these two contrasting sampling strategies lend further credence to the validity of their results. While the question ‘is RA genetic?’ is unanswered by the Danish data, we are surprised that the BMJ chooses to ask this question at all. Readers need only to consider the classical example of phenylketonuria (where an inherited disease is expressed only in the context of specific environmental exposure) to realise the conceptual difficulty faced in answering this question (9). The genetic susceptibility to tuberculosis (which shows a greater concordance in MZ than DZ twins (10)) also highlights the questionable relevance of attaching the label ‘genetic’ or ‘environmental’ to a complex disease. Estimating the relative contribution of genetic and environmental variation through twin data, rather than contributing to the century-old debate over ‘nature vs. nurture’, is of most value in providing a measure of the extent to which polymorphic variation in a population might account for the variation in occurrence of disease. These estimates give a measure of the likely success of specific study designs to identify individual genetic effects. The established importance of genetic variation in HLA in RA and the recent identification of new genetic regions linked to the disease (11) is ample evidence of an aetiological contribution from specific genes. (1) Svendsen AJ, Holm NV, Kyvik K, Petersen PH, Junker P. Relative importance of genetic effects in rheumatoid arthritis: historical cohort study of Danish nationwide twin population. BMJ 2002; 324(7332):264. (2) MacGregor AJ, Snieder H, Rigby AS, Koskenvuo M, Kaprio J, Aho K et al. Characterizing the quantitative genetic contribution to rheumatoid arthritis using data from twins. Arthritis Rheum 2000; 43(1):30-7. (3) Silman AJ, Hochberg MC. Epidemiology of the rheumatic diseases. Second ed. Oxford: Oxford University Press, 2001. (4) Jarvinen P, Koskenvuo M, Koskimies S, Kotaniemi K, Aho K. Rheumatoid arthritis in identical twins: a clinical and immunogenetic study of eight concordant pairs derived from a nationwide twin panel. Scand J Rheumatol 1991; 20(3):159-164. (5) MacGregor AJ, Bamber S, Carthy D, Vencovsky J, Mageed, RA et al. Heterogeneity of disease phenotype in monozygotic twins concordant for rheumatoid arthritis. British Journal of Rheumatology 1995; 34(3):215-220. (6) Aho K, Koskenvuo M, Tuominen J, Kaprio J. Occurrence of rheumatoid arthritis in a nationwide series of twins. Journal of Rheumatology 1986; 13(5):899-902. (7) Silman AJ, MacGregor AJ, Thomson W, Holligan S, Carthy, Farhan A et al. Twin concordance rates for rheumatoid arthritis: results from a nationwide study. British Journal of Rheumatology 1993; 32(10):903-907. (8) Jarvinen P. Occurrence of ankylosing spondylitis in a nationwide series of twins. Arthritis Rheum 1995; 38(3):381-383. (9) Khoury MJ, Thrasher JF, Burke W, Gettig EA, Fridinger F, Jackson R. Challenges in communicating genetics: a public health approach. Genet Med 2000; 2(3):198-202. (10) Bellamy R. Genetics and pulmonary medicine. 3. Genetic susceptibility to tuberculosis in human populations. Thorax 1998; 53(7):588-593. (11) Gregersen PK. Genetics of rheumatoid arthritis: confronting complexity. Arthritis Res 1999; 1(1):37-44. |
|||
|
|
|||
|
Charles M. Grossman, MD 1015 NW 22nd Avenue, Portland, Oregon 97210
Send response to journal:
|
The comment of MacGregor et al1 on your cover statement of 2 Feb., "Rheumatoid Arthritis: Is it Genetic?" and answer "probably not", do not really criticize your statement. I would consider it correct.2 Perhaps as purists, they might prefer the use of "non-genetic" instead of "environmental". To many patients genetic really means familial, and your comment might have used familial. The presence of some genes, which may contribute to the development of disease, does not make that disease familial. The same argument applies to cancer. It has been suggested that the "overwhelming contributor to the causation of cancer in the population of twins studied was the environment".3 It is well known that a number of genes have been identified as predisposing to non-familial cancers. Too little money has been spent on environmental contributions to cancer and other diseases. As an example, the U.S. "war on cancer" has cost many billions of dollars in the past 30 years4. Almost all has gone into molecular biology and genetics with tremendous progress in those fields. But, are we much better off in treating our patients with cancer than we were in 1975? Sincerely, Charles M. Grossman, MD
1 BMJ, Cover page, 2 Feb 2002 2 MacGregor AJ, Lanchbury J, Rigby AS, Kaprio J, Snieder H. BMJ (LeMer). 2002;324:1100-1. 3 Lichtenstein P, Holm NV, Verkasalo PK, et al. Enviornmental and hertable factors in the causation of cancer "Analyses of cohorts of twins from Sweden, Denmark, and Finland. The New England Journal of Medicine 2000;343(2):78-85. 4 Epstein SS. Legislative proposals for reversing the cancer epidemic and controlling run-away industrial technologies. Int J Health Serv 2000;30(2):353-71 |
|||
|
|
|||
|
Anders Jørgen Svendsen, Consultant Department of Rheumatology, Odense University Hospital, 5000 Odense, Denmark, Niels V. Holm, Kirsten Kyvik, Per Hyltoft Petersen and Peter Junker
Send response to journal:
|
The electronic response to our RA twin study by MacGregor et al is focused on three major points: the heritability estimate, potential bias and current evidence on the contribution by specific genes to the etiology of RA. We agree with MacGregor et al. that current evidence including our own does not allow a definite answer to the BMJ front page. However, we feel, that our data which are based on a stringent study design, implies that future research on the aetiopathogenesis of RA should pay more attention towards non-genetic factors than hitherto. MacGregor and al. state that based on standard modeling approaches, two previous twin studies on RA have reported comparable heritability estimates at approximately 60%. However, some important assumptions should be considered when calculating heritability, particularly when dealing with discrete, in this case binary, data such as RA versus non-RA. Heritability models for binary outcomes assume the existence of a normally distributed graded attribute referred to as the individual’s liability to the disease. Thus, the application of methods to assess heritability excludes situations where the variation is discontinuous, which would apply to diseases determined by a single major gene. Concerning RA only one possible genetic factor, the so called shared epitope, has been identified and confirmed in several epidemiological as well as segregation studies.1-2 Thus there is no definite evidence, that RA is a polygenic disease. Besides, it is often stated that the HLA component accounts for 37% of the genetic contribution.3 Hence, if a heritability figure of 60% is correct, then the assumption of no major genetic contribution would appear to be violated. The heritability estimate, therefore, is based on a theoretical concept relying on several assumptions which cannot at present be accounted for in RA. Besides, heritability estimates are population specific since the variation of environmental and genetic effects can not be assumed to be identical in various geographic areas and ethnic populations. In our study, twins below 42 years of age covered 92% of the total population at risk thereby skewing the sample towards the young age groups where it is well established that the prevalence of RA is low. The prevalence of RA in our study accords well with the latest population based epidemiologic study from Norway where the overall prevalence of RA between 20 and 79 years was 0.437 (95% CI 0,413 – 0,461) and prevalences exceeding 1% were only found beyond 60 years of age.4 With regard to the internal validity of our study, Macgregor and al. states that the concordance estimate of zero is by definition biased since MZ concordant pairs do exist as reported in previous twin studies on RA.5-6 Bias is a systematic error originating from problems inherent in a study including selection, information or confounding. In our study all possible measures have been taken to minimize bias e.g. by contacting all available individuals in the national Danish twin population irrespective of disease status, by using multiple sources of information and by doing specialist examination of all candidate probands. Furthermore, it can be questioned whether concordant pairs actually do exist in excess of the coincidence rate. In the Danish population approximately 2 MZ and 3 DZ concordant pairs are expected by chance and approximately the same number would be anticipated in Finland based on a point prevalence of RA at 1%, a twining rate at 1% and a population at risk of 5 million. In the Finnish twin study, an excess number of concordant pairs was identified, 8 MZ and 6 DZ concordant pairs. The fact that an excess number of concordant twins were retrieved in Finland is interesting but does not imply that a similar number would be identified in Denmark. Concerning the UK study 19 MZ and 39 DZ twin pairs would be anticipated to be concordant by chance based on the same assumptions. That study, however, identified 14 MZ and only 4 DZ concordant pairs and hence no concordant MZ pairs in excess of what could be expected by chance. Besides, in a further effort to minimize the risk of missing possible concordant pairs we used several alternative and independent retrieval sources. Therefore, we do not agree that our study is biased by definition, simply because we failed to identify MZ pairs concordant for RA and particularly not when all the available literature on twin methodology points to the risk of bias toward overascertainment of MZ pairs, and concordant MZ pairs in particular.7 MacGregor et al. argue that the established importance of genetic variation in HLA in RA and the recent identification of new genetic regions linked to the disease is ample evidence of an aetiological contribution from specific genes. However, in a community based incidence study of early RA no increased frequency of DR4 and DR1 was observed.8 On the other hand, early RA carriers of the shared epitope may be at particular risk of developing erosions and persistent disease .9-11 Thus, the HLA component is not a prerequisite for developing RA but is rather a contributing genetic determinant for the disease severity. If, however, disease severity is genetically determined then one would expect that MZ twin pairs concordant for RA were homogenous with respect to clinical expression of the disease, which does not appear to be the case.12 Thus, the finding of several possible genetic risk factors is interesting but their significance needs to be confirmed just as the bulk of putative environmental effectors proposed so far.13 Experience from type 1 diabetes, which based on twin studies has a larger genetic contribution, suggests that this will be a tedious process. We would like to point out that upon initiation of the Danish RA twin study we were if anything biased towards finding a genetic contribution to the etiology of RA. With a balanced view to the strengths and weaknesses of previous RA twin studies we find it justified to report our discordant results which are based on carefully designed proband retrieval among complete age segments of all Danish twins irrespective of disease state. We believe that the RA syndrome is most likely a disease with a heterogeneous etiopathogenesis and that twin studies are the best tool within the field of genetic epidemiology to address the question of the relative importance of genetic and environmental factors in the disease process. In order to reduce bias we suggest the following requirements in future twin studies on RA: • External validity is estimated by measuring the occurrence of the disease in the twin population and by comparing it with the occurrence in the general population. • Internal validity is estimated by comparing disease characteristics and discordance time of both zygosity groups. • Completeness of the study is estimated by using at least two independent recruitment sources. • Identification of cases as probands and secondary cases and use of the probandwise concordance rate which is the most unbiased concordance rate under incomplete ascertainment and which is comparable to the recurrence risk in other groups of relatives • Validation of the diagnosis according to contemporary classification criteria used in other epidemiological and clinical studies. There is no conflict of interest. Reference List 1. Rigby AS. HLA haplotype sharing in rheumatoid arthritis sibships: risk estimates in siblings. Scand.J.Rheumatol. 1992;21:68-73. 2. Tiwari JL, Terasaki.P.I. HLA and disease associations. New York: Springer Verlag, 1985. 3. Deighton CM, Walker DJ, Griffiths ID, Roberts DF. The contribution of HLA to rheumatoid arthritis. Clin.Genet. 1989;36:178-82. 4. Kvien TK, Glennas A, Knudsrod OG, Smedstad LM, Mowinckel P, Forre O. The prevalence and severity of rheumatoid arthritis in Oslo. Results from a county register and a population survey. Scand.J.Rheumatol. 1997;26:412-8. 5. Aho K, Koskenvuo M, Tuominen J, Kaprio J. Occurrence of rheumatoid arthritis in a nationwide series of twins. J.Rheumatol. 1986;13:899-902. 6. Silman AJ, MacGregor AJ, Thomson W, Holligan S, Carthy D, Farhan A et al. Twin concordance rates for rheumatoid arthritis: results from a nationwide study. Br.J.Rheumatol 1993;32:903-7. 7. Lykken DT, Tellegen A, DeRubeis R. Volunteer bias in twin research: the rule of two-thirds. Soc.Biol. 1978;25:1-9. 8. Thomson W, Pepper L, Payton A, Carthy D, Scott D, Ollier W et al. Absence of an association between HLA-DRB1*04 and rheumatoid arthritis in newly diagnosed cases from the community. Ann.Rheum.Dis. 1993;52:539-41. 9. Gough A, Faint J, Salmon M, Hassell A, Wordsworth P, Pilling D et al. Genetic typing of patients with inflammatory arthritis at presentation can be used to predict outcome [see comments]. Arthritis Rheum. 1994;37:1166-70. 10. Salmon M, Wordsworth P, Emery P, Tunn E, Bacon PA, Bell JI. The association of HLA DR beta alleles with self-limiting and persistent forms of early symmetrical polyarthritis. Br.J.Rheumatol. 1993;32:628-30. 11. Heard R. HLA and autoimmune disease. In Robert Lechler, ed. HLA & DISEASE, pp 136-8. London NW1 7DX: Academic Press Limited, 1994. 12. MacGregor AJ, Bamber S, Carthy D, Vencovsky J, Mageed RA, Ollier WE et al. Heterogeneity of disease phenotype in monozygotic twins concordant for rheumatoid arthritis. Br.J.Rheumatol. 1995;34:215-20. 13. Silman AJ, Hochberg MC. Epidemiology of the Rheumatic Diseases. New York: Oxford University Press Inc., 1993. |
|||
|
|
|||
|
dr.manan vasenwala, consultant-cardiologist k.k.heart center, aligarh-202002.india
Send response to journal:
|
one common cause of multifactorial diseases is stress. the incidence of rheumatoid arthritis was thought to be less in india due to the understanding that it is a disorder of temperate regions. in general practice,in the past, it was rare to come across full blown case of arthritis. however, the trend is changing. i am beginning to see large number of cases. most of them are precipitated by a stressful event like divorce, or husband remarrying or custodial wrangle over a male child. other stressful events like death of parents etc also is common. in addition, a plethora of other auto-immune disorders like scleroderma, wageners' granulomatosis, sle etc have also cropped up.i suspect enviromental changes of which mainly stress is responsible.another possibility is enviromental pollution.of course, availbility of diagnostic techniques may be a factor, but not entirely. Competing interests: None declared |
|||