NEUROLOGY 1987;37:717-719 Louise S. Gresham, MPH; Craig A. Molgaard, PhD, MPH; Amanda L. Golbeck, PhD; and Richard Smith, MD For several case-control studies, fractures and me- chanical injuries were more frequent among amyotrophic lateral sclerosis (ALS) patients than con- trols. In the case-control study of Campbell et al, I ALS patients displayed 14% more frequent history of frac- ture; 13% of the cases were diagnosed with disorders of the axial skeleton, compared with 5.4% of the controls. Kurtzke and Beebe,2 using military records of 504 men who died of ALS, found excess hospital admissions for trauma and fracture, particularly of the limbs and skull. In a study of antecedent events, Kondo and Tsubaki3 found that mechanical injuries were two to three times more frequent in ALS patients, frequently occurring within 5 years of onset. Our case-control study4 of SUS- pected risk factors of ALS included assessment of skel- etal fractures. Methods. This study was carried out from January to May 1985. Sampling was done from a case listing developed from the Center for Neurologic Study, a patient support and research facility in San Diego. In addition, letters were sent to area neurologists detail- Table 1. Frequency of exposure to fracture among case-control pairs + + + - - + OR Males 6 9 10 8 0.90 0.30 2.67 Females 13 6 7 6 0.86 0.04 3.35 All 19 15 17 14 0.88 0.39 1.94 + + Both ca-,e and control exposed. - Case exposed, control not exposed. + Case not exposed, control exposed. - Both case and control not exposed. Table 2. Number of persons with fractures by site: Table 3. Average time from fracture to diagnosis, and ALS cases, controls* diagnosis to fracture among ALS cases ALS cases Controls All Males Females Site Male Female Total Male Female Total Head 0 0 0 3 1 4 Average number of years 19.5 (14-8) 24.5 (1:3.5) 15.3 (15.0) Neck 0 0 0 0 1 1 from fracture to diag- Spine 1 2 :3 1 :1 4 nosis Extrem- 1 4 2 1 35 13 20 33 Average number of years @3.0 (1.8) ;3.8 (2.6) 2.5 (1.0) ities from diagnosis to fracture 'I'otal 38 42 Standard deviation. One I)erst)n can have more than one fracture. ing the goals of the study and seeking to increase case identification. Identification of cases. (1) Objective diagnostic crite- ria: clinical picture consisting of (a) progressive mus- cular atrophy and weakness, (b) fasciculation, (c)corticospinal tract signs, and (d) no sensory loss. (2)Eligibility criteria: incident cases diagnosed during 1975-1985. Cases included lower and upper motor neu- ron involvement. Definition of controls. Friends or neighbors of ALS patients, matched one-to-one on age (ñ5 years) and sex, not former co-workers, and free of known neu- rologic disease. Data collection. Data collection was achieved by means of a mailed, self-administered questionnaire. Sixty ALS cases and 15 surrogates for ALS patients responded; .66 were used after screening for com- pleteness and data quality. There were 33 men (mean age at diagnosis, 55.2) and 33 women (mean age at diagnosis, 57.7). Ninety-seven percent of the cases and controls were white. Each ALS case was-given an addi- tional questionnaire to be completed by a friend or neighbor of similar age. Subjects completing the ques- tionnaire responded to demographic items and history of skeletal fracture, date of occurrence, and site of frac- ture. Analysis. Retaining matching, McNemar's test was used to test the hypothesis that the odds ratio (OR) was unity for history of having at least one fracture. Ninety- five percent approximate confidence intervals (CI) were constructed.5 Results. In comparing the number of individuals who reported at least one fracture, there was no significant difference in ALS cases and controls. Using McNemar's formula for pair matching, the overall OR was 0.88, with a 95% CI of 0.39,1.94. Among men, the OR was 0.90 (Ci = 0.30,2.67), and among women it was 0.86 (Cl = 0.04, 3.35) (table 1). Thirty-four (52%) cases reported at least one frac- ture (15 men, 19 women), compared with 36 (55%) of the control group (16 men and 20 women). Of those cases having a history of skeletal fracture, the mean number of fractures was 1.40 (SD = 0.75), men 1.13 (SD = 0.35) and women 2.47 (SD = 3.39). Among controls, the mean was 1.5 (SD = 0.91), men 1.56 (SD = 0.89) and women 2.25 (SD = 3.59). Four percent of the ALS patients and 14% of the controls reported fractures involving the head, neck, or spine (table 2). Among cases, 68% of the fractures oc- curred before diagnosis. Twenty-three percent occurred within 5 years of diagnosis, 19% within 10 years, and 58% more than 10 years before diagnosis. The average number of years from the time of fracture to the time of diagnosis was 24.5 in men (SD = 13.5) and 15.3 in women (SD = 15.0). The average number of years from diagnosis to fracture was 3.8 for men (SD = 2.6) and 2.5 for women (SD = 1.0) (table 3). Discussion. A number of theories have been proposed concerning the etiology of ALS, especially fractures and disorders of the axial skeleton .13 Fractures and skeletal abnormalities may imply abnormal mineral metabo- ]ism. In particular, the role of Ca+ + homeostasis can be seen in hyperparathyroidism, in which a neuro- muscular syndrome may be similar to that of ALS. Metabolic abnormalities could prevent proper detox- ification of environmental toxins, such as heavy metals. The significance of skeletal demineralization in the pathogenesis of ALS is unclear; different metabolic or toxic factors could be at work. Other factors have also been proposed. In our investigation of antecedent events of ALS, there was no difference between cases and controls in the number of reported skeletal fractures. The control group reported three times more fractures of the head, neck, and spine region than did ALS cases. Limbs ac- counted for most fractures in both groups. General trauma was not assessed. Unlike the findings of Felmus et al,l only 16% of the fractures among ALS cases occurred within 5 years of diagnosis. Most were more than 10 years before diagno- sis. If susceptibility to an injury is an indication of the early stages of disease, most fractures should occur shortly before onset. Fractures after diagnosis occurred within an average of 3 years. As in the study of Kurtzke and Beebe,' there was no predilection for the head, neck, or spine region. In this population, the data did not suggest exposure to skeletal fracture as being a primary etiologic factor in the development of ALS or as serving to promote symptoms. From the Division of Epidemiology and Biostatiqties (Mg. Gresham and Dr. Molgaard), Graduate School of Public Health. and the Department i)f Mathe- matical Sciences (Dr. Golkwk), San Diego SLate University; and the Center for Neuri)logic Study (Dr. Smith), San Diego, CA. Received April 28, 1986. Accepted for publication in final form August 5, 1986. Address correspondence and reprint requests to M%. Gresham, Division of Epidemiology and Biostatistics, Graduate School i)f Public Health, San Diegi) State University, San Diego, CA 92182@ References 1.Campbell AMG, Williams ER, Baritrop D. Motor'neuron disease and exposure to lead. i Neurol Neurosurg Psychiatry 1970;3:3:977-985. 2.Kurtzke JF, Beebe GW. Epidemiology i)f amyotrophic lateral scle- rosis: a case-control comparison based on ALS deaths. Neurology 1980;30:453-462. 3.Kondo K, Tsubaki'l'. Case-control studies of motor neuron disease: association with mechanical injuries. Arch Neurol 1981;38:220-226. 4.Gresham LS, Molgaajrd CA, Golbeck AL, Smith IIA. Amyotrophic lateral sclerosis and occupational heavy metal exposure: a case- control study. Netiroepidemioli)gy 1986;5:29-38. 5.,Johnson N, Kotz S. Distributions in statistics. Boston: Houghton Mifflin Co, 1969. 6.Pierce-Ruhland 1', Patten B. ltel)eat study of antecedent events in motor neuron disease. Ann Clin Res 1981; 13:102-107. 7.FeIMLIc; MT, Ilatten BM, Swaiike L. Antecedetit events in amytrophic lateral sclerosis. Neurology 1976;26:167-172.