Journal of Neurology, Neurosurgery, and Psychiatry 1992;55:997-1001 997 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.11.997 on 1 November 1992. Downloaded from The environment in childhood and risk of motor neuron disease C N Martyn, C Osmond Abstract motor neurons, such individuals would be To investigate a possible link between especially vulnerable to further neuronal loss, subclinical infection with poliovirus in perhaps through ageing or because of a second childhood and increased risk of motor insult to the nervous system. neuron disease in adult life, environmen- A prediction from this hypothesis is that risk tal determinants of infection in early life of motor neuron disease will be related to were compared in 98 cases of motor factors that determine likelihood of poliovirus neuron disease and 335 age and sex mat- infection in childhood and the age at which it ched controls. A weak but consistent rela- occurs. The prevalence of seropositivity to tion was found between motor neuron poliovirus in the middle aged and elderly is disease and factors in the childhood envi- high9 partly because exposure to wild type ronment known to increase likelihood of virus during the first half of the century was enteric infection. Relative risks associated common and partly as a result of programmes with spending the first 10 years of life in a of active immunisation against the virus. house without domestic amenities such as Unfortunately, serological techniques do not a bathroom, running hot water or flushing distinguish between those who acquired anti- lavatory, living in overcrowded condi- bodies as a result of infection with wild type tions, frequent changes of address or virus and those who are seropositive because having a sibling with paralytic poliomye- they have been immunised; they cannot there- litis were all greater than unity, although fore be used to estimate childhood exposure to only those for absence of running hot poliovirus retrospectively. Population-based water and frequent changes of address surveys carried out in the 1950s, 102 however, were statistically significant. before mass immunisation programmes were instituted, identified the important environ- (X Neurol Neurosurg Psychiatry 1992;55:997-1001) mental determinants of poliovirus infection as the presence or absence of domestic amenities such as running hot water, flushing lavatories In England and Wales the current geographical and bathrooms, degree of domestic over- distribution of mortality from motor neuron crowding and social class. disease correlates closely with the pattern of We present the results ofa case control study notifications ofpoliomyelitis in the past. ' There designed to investigate whether these variables http://jnnp.bmj.com/ are also similarities in the time trends of the were associated with risk of motor neuron two diseases. The steady rise in mortality from disease. motor neuron disease seen in Europe, Scandi- navia and North America over the past 25 years" has paralleled the increase in incidence Patients and methods of poliomyelitis that occurred in the same We attempted to recruit all prevalent cases of countries during the first half of the century. motor neuron disease resident in the area on September 23, 2021 by guest. Protected copyright. These epidemiological data suggest that there served by the Wessex Regional Health Author- might be a causal relationship between the two ity. To achieve as complete ascertainment of conditions. Although evidence from case series cases as possible, several different methods suggests that a history ofparalytic poliomyelitis were used to identify individuals with the can be obtained from patients with motor disease. Principal sources of information were neuron disease more often than could be the case records of the Wessex Neurological chance 6 is clear that General a MRC Environmental expected by alone,5 it Centre, Southampton Hospital, Epidemiology Unit, most patients cannot recall ever having had diagnostic register maintained by the clinical Southampton General poliomyelitis. Any link therefore, must be with neurophysiology department of the Wessex Hospital, subclinical infection with poliovirus rather Neurological Centre and a list of members of Southampton C N Martyn than with poliomyelitis itself. the Motor Neuron Disease Association whose C Osmond It is possible that subclinical forms of polio- addresses corresponded to the geographical " Correspondence to: virus infection exist in humans,' as have been area of the study. All patients included in the Dr C N Martyn, MRC shown in in which the infection as cases seen a Envirornmental Epidemiology primates,8 study had been by consultant Unit, Southampton General involves the central nervous system but is not neurologist and all diagnoses were checked by Hospital, Southampton S09 severe to cause muscle weakness. the case records for the 4XY, UK. enough reviewing original or, these subclinical cases would never few cases where this was not Received 10 September Although possible, by 1991 and in revised form have been diagnosed as having poliomyelitis, writing to the neurologist whose care the 20 January 1992 and motor neurons cri- 16 March 1992. considerable loss of could have patient was under. Minimum diagnostic Accepted 23 March 1992 occurred. Left with a depleted population of teria were clinical or electrophysiological signs 998 Martyn, Osmond J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.55.11.997 on 1 November 1992. Downloaded from of both upper and lower motor neuron disease the recorded address was no longer current or and that both brainstem and at least one region because they had died. Of the 502 who were of the spinal cord were affected. In addition, sent a letter explaining the study and request- evidence was required ofprogression of disease ing permission for an interviewer to visit, 167 over time. Cases who showed atypical clinical declined to take part. Interviews were obtained features or in whom the diagnosis was doubtful from 335 subjects-a response rate of 67%. were not entered into the study. To avoid any For three cases we did not succeed in inter- confusion with cases of late degeneration after viewing three or more matched controls; each poliomyelitis, any subject whose medical of these cases was matched with two controls. records indicated that they had suffered from Three controls from whom interviews had poliomyelitis in the past or who, when inter- been obtained were excluded from the analysis viewed, gave a history of acute paralytic polio- because they were matched to the case who, myelitis was excluded. during the interview, gave a history of polio- In the later stages of the study, when it myelitis. became clear that insufficient numbers of cases Demographic characteristics of the cases in could be identified within the area defined this study were fairly typical of patients with above, we recruited patients from contiguous motor neuron disease in the UK: 61% were district health authorities outside the Wessex men; median age at the time of interview was health region who were under the care of 62 years (range 32-84); and all were white. No neurologists in Bristol, Gloucester and Oxford. case gave a family history of motor neuron The same criteria for diagnosis were applied. disease. For each case, six potential controls, mat- The method used for matching controls by ched for sex and within the same five year age age proved satisfactory; 86% of controls were band as the case (40-44, 45-49, 50-54 etc), within two years of age of the case and there were selected from the list of the general was no systematic tendency for controls to be practitioner. who had originally referred the older or younger than cases. case for investigation. The list of patients Analysis of the data was carried out using registered with the general practitioner was conditional logistic regression techniques to searched forward from the entry for the case take account of the matching of cases and and the first six eligible controls were identi- controls.'3 Results have generally been pre- fied. sented as an odds ratio as an estimate of After obtaining permission from the general relative risk. practitioner, patients and the first four controls selected for each case, were sent a letter requesting permission for an interviewer to Results call. Those who failed to respond were con- There was a consistent relation between both tacted by a personal visit. Where fewer than the absence of domestic amenities and over- three controls agreed to be interviewed, the crowding within the home during the first 10 remaining two controls were approached in the years of life and risk of motor neuron disease. same way. The interviewer administered a Odds ratios associated with the lack of a room questionnaire enquiring about the subject's with a fixed bath, lack of a running hot water domestic, social and family circumstances system, lack of an indoor flushing lavatory and during the first 10 years of their life. All domestic crowding of 1 5 persons per bedroom http://jnnp.bmj.com/ interviews, both of patients and controls, were or more were all greater than unity when carried out by the same person. compared with the presence of these amenities From the various sources described above, or a lesser degree of crowding. None of these 223 possible cases of motor neuron disease odds ratios was large and only that for lack of a were identified. Of these, 77 had died by the hot water system was statistically significa'nt. time we made contact with the general practi- These results are shown in table 1. tioner. In 21 cases the was either not No relation was between social diagnosis found class on September 23, 2021 by guest. Protected copyright. confirmed as motor neuron disease or there at birth (assessed by father's occupation), size was an indication in the case records that the of family, educational level (assessed by age at patient had previously suffered an attack of which the subject ceased full time education), acute paralytic poliomyelitis.