Is disabling fatigue in childhood influenced by genes?

Psychological Medicine, 1999, 29, 279–282. # 1999 Cambridge University Press Printed in the United Kingdom Is disabling fatigue in childhood influenced by genes ? A. F A R M E R," J. S C O U R F I E L D, N. M A R T I N, A. C A R D N O    P. M  G U F F I N From the Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, London ; and Department of Psychological Medicine, University of Wales College of Medicine, Cardiff ABSTRACT Background. Medically unexplained chronic fatigue in childhood may cause considerable disability and (by definition) its cause remains unclear. A study of fatigue in healthy twins has been undertaken to examine whether or not genetic factors play a part. Method. A questionnaire survey of the main carers of an epidemiological population-based sample of 670 twin pairs who were asked about periods of unexplained and disabling fatigue in their twins. Out of 1340 individuals a period of disabling fatigue was reported for 92 (6±9 %). Thirty-three (2±5 %) reported disabling fatigue for more than 1 month. Zygosity could be confidently assigned in 98 % of the sample providing 278 monozygotic (MZ) and 378 dizygotic (DZ) pairs. These data were analysed using a structural equation modelling approach. Results. The results showed that disabling fatigue in childhood is highly familial with an MZ tetrachoric correlation (rMZ) of 0±81 and a DZ tetrachoric correlation (rDZ) of 0±59, for fatigue lasting at least a week. The most acceptable model using Akaike’s information criteria, was one containing additive genetic effects (A) and shared environment (C) plus residual (or non-shared) environment (E). For fatigue lasting at least a month rMZ was 0±75 and rDZ 0±47. The most acceptable model included just A and E. However, the role of shared environment could not be conclusively rejected. Conclusions. Unexplained disabling fatigue in childhood is substantially familial. Both genetic and shared environmental factors are worth further exploration in a search for the causes. INTRODUCTION Persistent fatigue, unexplained by obvious medical disorder occurs in children and adolescents although less research has been undertaken in these age groups compared to adults (Joint Working Group of Royal Colleges of Physicians, Psychiatrists and General Practitioners, 1996). There is little information on the prevalence or other aspects of the epidemiology of fatigue in children. It has been suggested (Joint Working Group of Royal Colleges of Physicians, Psychiatrists and General Practitioners, 1996) that the research that has been carried out has emphasized the psychosocial aspects of chronic " Address for correspondence : Professor A. E. Farmer, Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, 111 Denmark Hill, London SE5 8AF. fatigue in somewhat selected samples. However, chronic fatigue in children and adolescents may still be quite disabling and it has been claimed that it is the single commonest cause of longterm absence from school in the UK (Dowsett & Colby, 1997). There has been considerable debate about the aetiology of chronic fatigue and much of this has centred on the role of persistent viral infection or psychosocial stressors or a combination of the two (Pawlinowska et al. 1994). In addition, we are aware that among those who assess or treat such subjects there is the clinical impression that persisting fatigue clusters in families. Although there is not, to our knowledge any hard evidence to support such a clinical impression it is worthy of further exploration for three reasons. First, infections tend to run in families. Secondly, genetic factors may con- 279 280 A. Farmer and others tribute to vulnerabilities. Lastly, attitudes, beliefs and aspects of culture are also ‘ transmitted ’ within family members and for example, this could include ‘ abnormal illness ’ behaviour, which has been suggested as being important in the presentation of symptoms of fatigue (David et al. 1988). We, therefore, set out to explore whether symptoms of fatigue are familial in a large epidemiological sample of school aged twins. METHOD A sample of school-aged twins (aged 5–17 years) was systematically ascertained in the district served by the Bro Taf Health Authority, South Wales. A total of 1109 twins was surveyed by mailed questionnaires that were completed by their parents. This represents the total number of twins identified via community paediatric registers who were within the age range, were both still alive and where the current address was traceable. Zygosity was determined by a reliable questionnaire as previously described (Thapar & McGuffin, 1997). The parents were asked to report whether either member of their twin pair had ever in their life had a period of a week or more when the child felt very tired and unable to do his}her usual activities, such as going to school or college or out to enjoy themselves with friends. If such an episode had occurred, further questions were asked about the duration of tiredness. Fatigue was considered as a threshold trait, liability to which is normally, or approximately normally, distributed in the population with only those whose liability exceeds a certain threshold being classed as affected (Falconer, 1965). The threshold was arbitrarily defined as 1 week or more of disabling fatigue for the first set of analyses and 1 month or more for the second set. Simple univariate analyses were carried out based on the observed tetrachoric correlation for MZ and DZ pairs with the aim of arriving at the most parsimonious explanation of the data. This was done by comparing a full model where the phenotype is contributed to by additive genetic (A), common (shared family) environment (C) and unique (or residual) environmental effects (E), with reduced models (McGuffin et al. 1994 ; Neale, 1997), using Akaike’s information criteria (AIC) and chi-square differences. RESULTS Completed questionnaires were received on both members of 670 pairs representing a response rate of 60 %. Since it has been suggested that there may be an association between socioeconomic status and symptoms of fatigue (Joint Working Group of Royal Colleges of Physicians, Psychiatrists and General Practitioners, 1996) we checked for bias in response. There were no significant differences between those who replied and those who did not with respect to ‘ ecological ’ sociodemographic indices that could be obtained by comparing the postcodes of responders and non-responders (Townsend & Davidson, 1992). In 14 pairs the zygosity information was equivocal and these were excluded from analyses leaving a total of 278 monozygotic (MZ) and 378 dizygotic (DZ) pairs (180 same sex and 198 opposite sex). Of the entire sample of 1340 individuals, a period of disabling fatigue was reported for 92 (6±9 %) and 33 (2±5 %) had had a period of disabling fatigue lasting more than a month. Among other measures obtained on this sample were responses on the Strengths and Difficulties Questionnaire (Goodman, 1997), which provides an approximate but convenient guide to psychiatric ‘ caseness ’. Caseness on this measure was significantly associated in the total sample with disabling fatigue lasting a week (cases in twins with fatigue ¯ 27 %, cases in twins without fatigue ¯ 11 %, χ# ¯ 16±76 degrees of freedom (df) ¯ 1, P ! 0±0001). Similarly, disabling fatigue for at least a week was also significantly associated with parents stating that the twin found it ‘ hard to get to sleep ’ (χ# ¯ 56±0, df ¯ 1, P ! 0±0001) and the twin having migraine (χ# ¯ 16±7, df ¯ 1, P ! 0±0001). When the twins were divided into those under 12 and those who were 12 years and older, the older group were significantly more likely to have had a week or more of disabling fatigue (χ# ¯ 9±6, df ¯ 1, P ! 0±002). The data were analysed using a structural equation modelling approach as implemented in the Mx program (Neale, 1997). There were no significant differences in the reported rates of fatigue in boys or girls and no greater similarities A twin study of fatigue in childhood Table 1. Components of variance in liability to fatigue lasting at least 1 week Model ACE AE CE E A C E χ# df P AIC 0±42 0±83 [0] [0] 0±38 [0] 0±7 [0] 0±20 0±17 0±30 1±0 0±79 2±83 3±09 1±0 3 4 4 5 0±85 0±59 0±54 0±00 ®5±21 ®5±17 ®4±91 [0], parameter fixed at zero. AIC, Akaike’s information criteria. Table 2. Components of variance in liability to fatigue lasting at least 1 month Model ACE AE CE E A C 0±54 0±76 [0] [0] 0±19 [0] 0±58 [0] E χ# 0±26 1±3 0±24 1±505 0±42 2±36 [0] 54±56 P df 3 4 4 5 AIC 0±73 ®4±7 0±83 ®6±495 0±67 ®5±64 ! 0±0001 44±56 [0], parameter fixed at zero. AIC, Akaike’s information criteria. in same sex than in opposite sex as pairs. Taking the threshold as fatigue lasting at least a week, the tetrachoric correlations were 0±81 (95 % CI ¯ 0±61–0±92) for MZ and 0±58 (CI ¯ 0±33–0±78) for DZ pairs. An ACE model, allowing for additive genetic effects (A) common environmental effects (C) and non-shared or residual environmental effects (E) provided the best explanation of the data on the basis of the AIC (χ# ¯ 0±79, df ¯ 3, P ¯ 0±85) (Table 1). However, reduced models in which there was assumed to be either no shared environmental effect (AE) or no genetic contribution (CE) could not be conclusively rejected on the basis of chi-square difference or goodnessof-fit. A model assuming no familial effect (E) provided a very poor fit (χ# ¯ 54±75, df ¯ 5, P ! 0±0001). Taking the threshold as fatigue lasting at least 1 month, the pattern of findings was somewhat similar (Table 2). The MZ correlation was 0±75 (CI ¯ 0±28–0±94) and the DZ correlation 0±47 (CI ¯ 0±04–0±77) but an AE model now gave the most acceptable fit on the basis of the AIC. However, the difference in chi-square for reduced AE or CE models was again modest compared with the ACE model and a model assuming no familial effect provided a very poor fit (χ# ¯ 54±56, df ¯ 5, P ! 0±0001). 281 DISCUSSION The results suggest that the symptom of disabling fatigue in childhood is highly familial and that the source of familiality is likely to include genes acting additively. However, the role of shared environment, or of a combination of shared genes and shared environment, could not be rejected conclusively either for fatigue lasting at least a week or at least a month. Although, previous considerations of the causes of disabling fatigue have included the search for different viruses and speculation about sociocultural factors that could conceivably be familial, the possible role of genetic factors in the aetiology of the disorder has until now received scant attention. The obvious main limitation of the study is that the data were collected by questionnaire and provide only a broad and imprecise notion of what is meant by ‘ fatigue ’. The symptom was defined over two fairly short time periods and it is a matter of conjecture whether it is actually related to chronic fatigue syndrome (CFS), either in childhood or adult life. However, one clue that it might be, is that migraine headaches, sleep difficulties and a measure of psychological disturbance were significantly associated with fatigue. These symptoms are comparable to those included in the operational criteria for CFS (Fukuda et al. 1994). Another limitation is that the rating was performed by parents, rather than subjects themselves, but there is no suggestion from work on an earlier Welsh twin sample that this introduces a bias in overestimating familiality, although information from different types of informant does appear to influence estimation of heritability (Thapar & McGuffin, 1996 ; Simonoff et al. 1998). In addition, the limited response rate which is inevitable when using a mail-out questionnaire survey could have led to some bias in reporting. However, there is unlikely to have been a bias in the socio-economic status of those not responding. As far as we are aware this report is the first to provide information on the familiality and heritability of disabling fatigue in childhood based on a large systematic sample. To date there has been much debate in this area but little hard evidence concerning aetiology. Despite its limitations our study indicates that familial 282 A. Farmer and others factors are worth further, more detailed, exploration. This study was supported by the Medical Research Council with Training Fellowships to Drs Scourfield and Cardno and a Studentship to Mr Martin. REFERENCES David, A. S., Wessely, S. & Pelosi, A. J. (1988). Post viral fatigue syndrome : time for a new approach. British Medical Journal 296, 696–699. Dowsett, E. G. & Colby, J. (1997). Long term sickness absence due to ME}CFS in UK schools : an epidemiological study with medical and educational implications. Journal of Chronic Fatigue Syndrome 3, 29–42. Falconer, D. S. (1965). The inheritance of liability to certain diseases, estimated from the incidence among relatives. Annals of Human Genetics 29, 51–76. Fukuda, K., Straus, S. E., Hickie, I., Sharpe, M. C., Dobbins, J. G., Komaroff, A., Schluederberg, A., Jones, J. F., Lloyd, A. R., Wessely, S., Gantz, N. M., Holmes, G. P., Buchwald, D., Abbey, S., Rest, J., Levy, J. A., Jolson, H., Peterson, D. L. & Vercoulen, J. H. M. M. (1994). The chronic fatigue syndrome : a comprehensive approach to its definition and study. Annals of Internal Medicine 121, 953–959. Goodman, R. (1997). The strengths and difficulties questionnaire : a research note. Journal of Child Psychology and Psychiatry 38, 581–586. Joint Working Group of Royal Colleges of Physicians, Psychiatrists and General Practitioners (1996). Chronic Fatigue Syndrome. Royal College of Physicians Publication Unit : London. McGuffin, P., Owen, M. J., O’Donovan, M. C., Thapar, A. & Gottesman, I. I. (1994). Seminars in Psychiatric Genetics. Royal College of Psychiatrists : London. Neale, M. C. (1997). Mx : Statistical Modelling. 4th edn. Dept of Psychiatry, Medical College of Virginia : Richmond, VA. Pawlikowska, T., Chalder, T., Hirsch, S. R., Wallace, P., Wright, D. J. M. & Wessely, S. C. (1994). Population based study of fatigue and psychological distress. British Medical Journal 693, 763–766. Simonoff, E., Pickles, A., Hervas, A., Silberg, J. L., Rutter, M. & Eaves, L. (1998). Genetic influences on childhood hyperactivity : contrast effects imply parental rating bias, not sibling interaction. Psychological Medicine 28, 825–838. Thapar, A. & McGuffin, P. (1996). Genetic influences on life events in childhood. Psychological Medicine 26, 813–820. Thapar, A. & McGuffin, P. (1997). Anxiety and depressive symptoms in childhood – a genetic study of comorbidity. Journal of Child Psychology and Psychiatry 38, 651–656. Townsend, P. & Davidson, N. (eds) (1992). Inequalities in Health. The Black Report. Penguin Books : London.