Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

AN “OUTBREAK” OF JUVENILE DIABETES MELLITUS: CONSIDERATION OF A VIRAL ETIOLOGY

AN “OUTBREAK” OF JUVENILE DIABETES MELLITUS: CONSIDERATION OF A VIRAL ETIOLOGY AMERICAN JOURNAL OF EPIDEMIOLOGY Vol. 100, No. 4 Copyright © 1974 by The Johns Hopkins University Printed in U.SA. AN "OUTBREAK" OF JUVENILE DIABETES MELLITUS: CONSIDERATION OF A VIRAL ETIOLOGY J. CLARK HUFF, JOHN C. HLERHOLZER AND WILLIAM A. FARRIS Huff. J . C , J. C. Hierholzer (Respiratory Virology Section, Center for Disease Control. Atlanta, Ga. 30333) and W. A. Farris. An "outbreak" of juvenile diabetes mellitus: consideration of a viral etiology. Am J Epidemiol 100:277-287 . 1974.—Nine of twelve cases of juvenile diabetes mellitus, representing an unusual geographic and temporal cluster, were investigated for evidence that a specific viral infection might be etiologically related to their occurrence. Eight diabetics had experienced recent "viral-like" illnesses, predominantly repiratory, but these illnesses bore no uniform temporal relation to their onsets of diabetes. Diabetics demonstrated no serologic evidence of a recent viral illness common to all. Elevated titers to only one virus, coxsackie B-3, were more prevalent in diabetics than in controls (33% vs. 6%), but geometric mean titers of diabetics to a panel of 26 common viral antigens were similar to those of controls. These data neither support nor negate the hypothesis that infection with a specific virus precipitates juvenile diabetes mellitus. diabetes mellitus, juvenile; respiratory tract infections INTRODUCTION co-workers (6, 7) have provided experimen- tal evidence that mice infected with a Recent experimental and epidemiologic strain of encephalomyocarditis (EMC) data have generated a growing interest in virus may develop damage of the /9-cells of the possibility that viral infections may be the pancreatic islets and a resultant "dia- involved in the pathogenesis of diabetes betic-like" state. Furthermore, Gamble mellitus. Scattered case reports have sug- and co-workers (8) have reported that the gested that human diabetes may follow Coxsackie B viruses, picornaviruses similar such viral diseases as mumps (1-3) and to EMC virus, may be associated with congenital rubella (4, 5). Craighead and human diabetes. Gamble's data included higher prevalence of antibody to Coxsackie Received for publication February 26, 1974, and in B viruses in individuals with juvenile dia- final form May 28, 1974. 1 From the Viral Diseases Branch, Bureau of betes of recent onset than in controls or Epidemiology and the Virology Branch, Bureau of long-term diabetics. Laboratories, Center for Disease Control, Atlanta, Ga. In early 1972, a group of pediatricians in Abbreviations: CF, complement fixation; CMV, Pinellas County, Florida, noted a large cytomegalovirus; EMC, encephalomyocarditis; GMT, number of new cases of juvenile diabetes. geometric mean titer; HI, hemagglutination-inhibi- This report describes the results of an tion; IHA, indirect hemagglutination; MKC, monkey kidney cell; RSV, respiratory syncytial virus; SN, epidemiologic and laboratory investigation serum neutralization. of these cases. The authors gratefully acknowledge the support of Dr. J. P. Hanley, a pediatrician in Clearwater, Flor- BACKGROUND ida; Dr. G. M. Dame, Director, Pinellas County Health Department; and other physicians in Pinellas Pinellas County lies along the Gulf Coast County; and gratefully acknowledge the technical of Florida just north of the mouth of assistance of J. T. Cole, K. Hilliard, and W. B. Yarbrough. Tampa Bay. The flat lowlands of the 277 278 HUFF, HIERHOLZER, FARRIS county are covered by contiguous urban positive viral serologic tests from Pinellas and suburban areas, where 96 per cent of County were also reviewed for the same the county's half million residents live period. (1970 census). The largest cities are St. Laboratory. In mid-March 1972, rectal Petersburg (population 216,232) and Clear- and throat swabs were taken from newly water (population 52,074). Tourism is the diabetic children and other members of main industry in the county. Socioeconom- their households. Specimens placed in ically, most residents are middle class. tryptose phosphate broth with 0.5 per cent Because the county is situated on a penin- gelatin were frozen in Dry Ice. In the sula and excellent medical care is available laboratory, specimens were processed for in the county, Pinellas County residents inoculation onto two strains of human generally seek medical care within Pinellas diploid lung fibroblast (RU-1 and WI-38), County. one line of human epithelium (HEp-2), and primary rhesus monkey kidney cell METHODS (MKC) cultures, as previously described (9). The specimens were also inoculated Epidemiologic. Medical records at the into 10-day embryonated eggs (10). Each seven hospitals in Pinellas County where specimen was passaged at least three times children are usually admitted were re- for a total incubation period of four weeks. viewed. Records of individuals up to age 14 who were admitted between January 1968 Single (convalescent) serum specimens and March 1972 as cases of diabetes mel- were collected by venipuncture in mid- litus were examined. A new "case" of March from newly diabetic children, mem- juvenile diabetes was defined as an illness bers of their households, and a control 1) diagnosed as diabetes in a child not group of children from Pinellas County previously known to have the disease, 2) with no history of diabetes. Non-diabetic characterized by glycosuria and hyper- controls were matched with the newly glycemia (a blood glucose of a least 200 diabetic children for age and sex. Serum mg/100 ml), and 3) treated with insulin at samples were stored at -30 C until tested. the time of hospital discharge. The "inci- The sera were tested for antibodies to the dence" of juvenile diabetes mellitus in the prototype strains of a large number of pediatric age group was estimated by plot- respiratory viruses by complement-fixation ting all new cases by month of hospitaliza- (CF), hemagglutination-inhibition (HI), tion. No effort was made to identify new indirect hemagglutination (IHA), or serum cases in older age groups or new cases in neutralization (SN) tests, as appropriate persons not hospitalized. for each particular virus. The sera were Investigation of cases. Diabetic children heat-inactivated (56 C, 30 minutes) for all hospitalized early in 1972 were investigated serologic tests. Antibodies to influenza A in detail. With the consent of the child's and B, parainfluenza 1, 2, and 3, the parents, a detailed medical history was "soluble" and "viral" antigens of mumps taken from each family member. Each virus, the group-specific hexon antigen of family was questioned concerning associa- adenovirus, respiratory syncytial virus tions between them and the families of (RSV), Mycoplasma pneumoniae, herpes other diabetic children. simplex 1, and cytomegalovirus (CMV) Review of public health records. Physi- were assayed by the standardized CF test cians' reports of viral diseases to the county with overnight fixation of 5 units of com- health department were examined for the plement (11). Antibodies to influenza C, time period November 1971 through March parainfluenza 4A and 4B, and coronavirus 1972. Records of positive viral cultures and OC 43 were assayed by the standardized HI "OUTBREAK" OF JUVENILE DIABETES MELLITUS test with .01 M phosphate-buffered saline case every 2 months. The maximum num- diluent and spectrophotometrically stand- ber in a consecutive 2-month period was ardized 0.4 per cent mammalian or 0.5 per four in 1971. In February and March 1972, cent avian red blood cells (12, 13). Hemag- 12 cases occurred. Based on an "expected" glutinating antigens of parainfluenza 4A rate of four new cases or fewer in a 2-month and 4B were prepared for use in th e HI test period, the occurrence of 12 cases was a as described by Killgore and Dowdle (14). very rare event (p < .001). Antibodies to herpes simplex 1 and CMV Investigation of cases. Nine of the 12 were determined by th e IHA procedures of patients hospitalized in February and Bernstein and Stewart (15, 16) and corona- March and their families agreed to partici- virus 229E antibodies were measured by a pate in the detailed case investigations. similar IHA test (17). Neutralizing anti- The nine children with newly diagnosed bodies to the six Coxsackie B viruses were diabetes included five males and four fe- assayed by SN tests in MKC tissue culture males, ages 4 to 13 (mean age 8.3). Al- (9). Neutralizing antibody titers to EMC though all nine had been hospitalized at virus (r+ variant), assayed by plaque neu- either of two hospitals in central Pinellas tralization tests in L-cell tissue culture, County, their places of residence were were generously provided by Dr. John E. scattered throughout the county. No dia- Craighead, University of Vermont. betic child and family was acquainted with any other child and family, and none were RESULTS associated in educational, religious, occu- Epidemiologic. Thirty-five cases of juve- pational, or recreational activities. No nile diabetes diagnosed in children be- medication or toxic exposure was common tween January 1968 and March 1972 are to the nine. presented by month of hospital admission A summary of data from case investiga- in figure 1. In the 48 months prior to tions is shown in table 1. Onset of diabetic January 1972, 23 new diabetics were hospi- symptoms preceded hospitalization by 1 to talized, an average "incidence" of one new 7 weeks (mean 3 weeks). Two factors com- 1O 8 - in in 3 6- 4 - 2 - JFMAMJ JASONDlJFMAMJ J ASONDlj FMAM J JASOND JFMAMJ J AS ONDljFM 1968 1969 1970 1971 1972 Month of Hospitalization FIGURE 1. Thirty-five cases of newly diagnosed juvenile diabetes mellitus, by month of hospital admission, Pinellas County, Florida, January 1, 1968-March 31, 1972. 280 HUFF, HIERHOLZER, FARRIS TABLE 1 Summary of epidemiologic data for nine new cases of juvenile diabetes Viral illness in 3 Date of months preceding hos- initial Date of pitalization h capital- Family history of symp- Patient Age Sex ization diabetes Date Type toms* 1 6 M 1/16* 2/1 Maternal great-aunt 12/29 URI.t otitis media 2 4 F 1/29 2/11 Paternal aunt, paternal 12/25 "influenza-like" great-grandmother illness 3 11 2/11 2/7 F 1/11 Maternal grandfather Pharyngitis 4 9 F 2/7 2/14 Maternal and paternal 1/15 Gastroenteritis cousins 6 M 2/16 Paternal grandmother 12/25 URI 5 12/25 6 10 M 2/8 2/22 None 2/18 URI 7 12 2/14 3/4 Maternal grandmother 3/4 ? URI, otitis media 8 13 F 2/18 3/9 Unknown (adopted) None None 4 M 2/24 3/9 Maternal grandmother 2/15 "influenza-like" illness * December 1971 to March 1972. t URI, upper respiratory infection. mon to the majority of the diabetics were a in Clearwater in December 1971, and influ- family history of diabetes and an illness in enza was recognized throughout the county the 3 months preceding hospitalization in January 1972. Outbreaks of other report- considered by the patient's physician to be able viral illnesses, including mumps and of possible viral etiology. Seven of the nine rubella, had not occurred. Influenza was families knew of relatives with diabetes, confirmed by isolation of A/Hong Kong/68 but further information concerning the (H3N2) strains and by positive serologic diabetic relatives was unavailable. Two of studies. Positive serologic tests were also the diabetic relatives were reported to have obtained for parainfluenza and adenovirus. had onset of disease in young adulthood; ECHO-11 and Coxsackie A-12 were iso- the others apparently were adult onset, lated from individual patients, and in ad- insulin-independent diabetics. dition, Coxsackie B-2 and ECHO-8 were Eight of the diabetics had histories of a isolated from sewage effluents. With the preceding illness of possible viral etiology. exception of influenza, however, norre of The etiology of the illnesses was not deter- these agents was known to be associated mined. Seven of the illnesses were respira- with significant morbidity. tory, ranging from simple upper respiratory Laboratory. Cultures of rectal and throat to febrile lower respiratory illnesses. In swabs taken from eight newly diabetic only four of the eight cases, however, did children and nine of their family members the illness precede the onset of diabetic yielded no virus isolates. At the time of symptoms. attempted isolation in mid-March, there Review of public health records. In a was no respiratory illness in any of the search for viral illnesses prevalent in Pinel- individuals tested; isolation attempts were las County prior to the cluster of cases of therefore made only on the chance that juvenile diabetes, reports to the county there might still be virus excretion from a health department of viral diseases were preceding illness. reviewed. An outbreak of measles occurred Single serum samples were taken from "OUTBREAK" OF JUVENILE DIABETES MELLTTUS 281 nine diabetic patients within 6 weeks after high titers of antibody to 15 viruses and to hospitalization and within 12 weeks after M. pneumoniae were quite similar in dia- onset of symptoms. Unfortunately, earlier betics and controls for all agents except sera for either the respiratory or diabetes Coxsackie B-3 (table 3). Three of nine illnesses were not available. Single serum diabetics (33 per cent) had significant specimens were also collected from 21 titers (> 1:80) to this virus, whereas only 2 family members (contacts) and from 36 of 36 matched, non-diabetic controls (6 per matched and 20 unmatched control chil- cent) had similar titers. This difference dren at this time (table 2). All of these sera were tested for antibody to 15 common TABLE 2 respiratory viruses and to M. pneumoniae. Age and sex characteristics of study population The antibody levels of the case, contact, Sex Age and matched control groups were tabu- Fe- Me- lated both as numbers of persons with Total Male male dian Range "significantly high" antibody titers (table Diabetics 9 5 4 9 3) and as geometric mean titers (GMT) 4-13 Family con- 21 10 26 3-46 (table 4). tacts "Significant" antibody titers were de- Controls, 36 20 16 9 ' 4-13* fined as those levels of antibody (serum matched Controls, 56 8 endpoint dilutions) which we considered, 23 33 4-13 total through diagnostic experience with the test procedures used, to be suggestive of recent * Diabetics and controls matched for age (within infection. The prevalence of significantly 1 year), sex, locality, and time of serum collection. TABLE 3 Prevalence of elevated antibody titers in diabetic patients, family members, and matched controls % with elevated titer (significant) 21 family 36 matched Virus Test titer" 9 patients members controlst Influenza A CF 1:32 11 14 19 Influenza B CF 1:32 0 0 Parainfluenza 1 CF 1:32 0 3 Parainfluenza 2 CF 1:32 0 3 Parainfluenza 3 CF 1:32 11 0 8 Mumps (soluble) CF 1:16 0 5 5 Mumps (viral) CF 1:32 0 0 Adenovirus CF 1:16 0 14 6 RSV CF 1:32 11 14 14 Herpes 1 CF 1:32 0 10 7 0 0 16 M. pneumoniae CF 1:32 0 10 4 Coxsackie B-l SN 1:80 Coxsackie B-2 SN 1:80 0 14 13 Coxsackie B-3 SN 33 6 1:80 10 Coxsackie B-4 SN 11 17 1:80" 38 Coxsackie B-5 SN 0 4 1:80 0 Coxsackie B-6 SN 1:80 0 0 * A titer of this level or higher is considered elevated and therefore may be an indication of recent infection (within approximately 5 months). See text for complete definition. t The frequency of elevated titers in the diabetic and control groups was not significantly different at the .05 level by Fisher's Exact Test for any virus except Coxsackie B-3 (P = .047). 282 HUFF, HIERHOLZER, FARRIS was of borderline statistical significance (p Results of GMT's to nine additional viral = .047) by Fisher's Exact Test. antigens for all diabetics, selected family members, and an unmatched group of Family members of the diabetics were a non -diabetic children are shown in table 5. diverse group of 12 adults and nine chil- As in table 4, antibody levels to most dren and therefore were not compared with viruses were typical of pediatric groups. the control group except in general terms. The GMT to CMV, which was somewhat Elevated titers to adenovirus, herpesvirus, high in diabetics, was nonetheless quite and Coxsackie virus B-4 were most preva- similar to that of the non-diabetic group lent among family members. and was not in an unusual range. The CF The serologic data expressed as GMT's and IHA tests for herpes and for CMV to 17 viral antigens did not reveal any measure different aspects of infection, and single agent to which diabetics possessed a neither test indicated that these viruses significantly higher mean titer (table 4). were associated with the respiratory ill- The GMT of diabetics to Coxsackie B-3 nesses of the diabetics before they were was almost two dilutions higher than con- hospitalized. trols, but statistically, the difference was not significant by the t-test (p = .083). The Antibody levels to the six Coxsackie B GMT's of family members, despite their viruses are presented for each of the nine age and sex diversity, were similar to those cases, together with the serologic identifi- of diabetics and controls for all agents ex- cation of other possible recent respiratory cept Coxsackie B-4; the GMT of family illnesses, to provide a summary of the most members to Coxsackie B-4 was approxi- salient serologic data (table 6). The pattern mately one dilution greater than the GMT seen in this summary is one of total ran- of diabetics and controls. domness. Virtually each case had serologic TABLE 4 Geometric mean serum antibody titers to 17 viral antigens in diabetic patients, family members, and controls GMT* 21 family 36controlB 56 controls Virus Test 9 patients members (matched only)f (total) Influenza A CF 6.86 8.84 6.92 7.38 CF Influenza B 4.00 4.56 4.16 4.16 CF 4.72 Paiainfluenz a 1 4.32 4.58 4.88 CF 5.21 5.32 Parainfluenz a 2 5.44 5.14 Parainfluenz a 3 CF 8.64 5.75 8.00 7.90 Mump s (soluble) CF 5.88 4.93 5.83 5.80 Mump s (viral) CF 5.88 4.56 8.32 8.31 Adenovirus CF 5.04 5.94 5.99 6.09 RSV CF 12.70 9.76 9.33 9.63 Herpes 1 CF 4.00 7.74 5.24 5.52 M . pneumoniae CF 4.67 5.75 8.64 8.83 Coxsackie B-l 5.00 7.70 5.52 5.56 SN Coxsackie B-2 SN 6.15 8.94 13.53 13.75 Coxsackie B-3 SN 23.04 9.61 8.70 8.92 Coxsackie B-4 SN 25.05 10.40 11.56 14.70 Coxsackie B-5 SN 5.63 5.79 5.74 6.30 SN 5.61 5.44 5.14 Coxsackie B-6 5.00 * A geometric mean titer (GMT) of 4.00 = < 1:8 for CF tests and of 5.00 - < 1:10 for SN tests. GMT's are listed as dilution factor of endpoint serum dilutions. t The GM T in th e diabetic and matched control groups was no t significantly different (P > .05) for any virus by the (-test. "OUTBREAK" OF JUVENILE DIABETES MELUTUS TABLE 5 Prevalence of "significant" antibody titers in diabetics and geometric mean serologic titers of diabetics, selected family members, and selected unmatched controls to 9 viral antigens Geometric mean titers* Prevalence of significant 10 family 24 Significant titers in 9 9 diabetics members controls Virus Test titer diabetics (%) (ages 4-13) (ages 5-23) (ages 4-13) Influenza C HI 1:128 0 13.72 19.70 NDf Parainfluenza 4A HI 1:128 0 13.39 13.00 ND Parainfluenza 4B HI 1:128 11 9.69 12.96 ND Herpes 1 IHA 1:2048 0 6.35 14.04 6.50 Cytomegalovirus CF 1:256 22 46.31 28.89 22.42 IHA 1:2048 11 21.82 12.96 26.90 Coronavirus OC43 HI 1:64 ND 0 11.91 14.60 Coronavirus 229E IHA 1:64 11 ND 9.83 8.58 Encephalomyocarditis PN§ 1:4 1.31| 0 l.OOt 1.351: • GMT of 4.00 - <1:8. GMT's are listed as dilution factor of endpoint serum dilution. t Not done. t GMT of 1.00 - <1:2. 5 PN - plaque neutralization. TABLE 6 Summary of Coxsackie B antibody titers and significant titers to other viruses for 9 new cases of juvenile diabetes mellitus SN antibody titers to Coisaclde virus* Case B-l B-2 B-3t B-5 B-6 viruses B-4t 1 0 0 160 0 0 None 0* 2 0 10 0 40 0 0 CMV 3 0 0 20 160 0 0 None 4 0 0 0 0 10 0 RSV, parainfluenza 4B 5 0 0 40 0 0 0 Coronavirus 229E 6 0 0 160 40 0 0 CMV 7 0 20 80 0 0 0 Influenza A 8 0 40 0 20 20 0 Parainfluenza 3 0 0 9 0 0 0 0 None * Serum neutralization (SN) titers listed as dilution factor of endpoint serum dilutions, t Among the 56 controls, 37 had SN titers to B-3 of < 1:10, 9 had titers of 1:10-1:20, 7 had 1:40-1:80, and 3 were >l:160; 35 had SN titers to B-4 of <l:10, 5 had 1:10-1:20, 12 had 1:40-1:80, and 4 were >l:160. evidence of infection with a different virus has not yet been clarified (18). Some inves- in the preceding four to six months. It is tigators have concluded that both genetic therefore apparent that no one of the and environmental determinants are in- Coxsackie B viruses nor any one of the 23 volved (19). Among the characteristics viruses tested for could account for this suggestive of environmental determinants "outbreak" of diabetes. is a seasonality in the onset of juvenile diabetes. Both a 1926 study of patients from the north central United States (20) DISCUSSION and a 1969 analysis of English diabetics Heredity has been accepted as the major (21) described a nadir in the incidence in determinant of the occurrence of diabetes late spring and early summer and a broad mellitus, but the exact mode of inheritance 284 HUFF, HIERHOLZER, FARRIS peak between late summer and early Attention has been focused therefore spring. Data collected in Pinellas County upon a similar group of picornaviruses, the also hinted of a seasonal pattern: 11 of the Coxsackie B viruses, which are extremely 23 cases identified from 1968 through 1971 frequent human pathogens. The Coxsackie (figure 1) were hospitalized in the months B viruses, like EMC, may cause pancreati- January-April, approximately the time of tis (27) and even diabetes (28) in experi- the 1972 peak. Many viral respiratory in- mental mice, and, in addition, they may fections have a similar seasonal pattern, attack the human pancreas in certain in- and that characteristic has suggested a fant infections (29). Although there are no possible cause of the seasonality of juvenile published case reports of diabetes in hu- diabetes. mans following proven Coxsackie B infec- tions, the epidemiologic and serologic data Although case reports have indicated of Gamble and co-workers (8) aroused that a diabetic state may follow such considerable interest in a possible relation diseases as mumps and congenital rubella between these viruses and human diabetes. (1-5), these reports have remained inciden- Gamble et al. reported that insulin- tal observations. There is, at present, no dependent diabetics within 3 months of firm evidence that one specific viral illness onset had higher serum neutralization ti- plays a significant part in the pathogenesis ters to Coxsackie B viruses, especially B-4, of human diabetes mellitus. than did non-diabetic controls or long- The most convincing data suggesting a standing diabetics (8). Controls in this viral etiology of diabetes come from the study, however, were not closely matched experiments carried out by Craighead and for age, sex, residence, or time of collection co-workers using a mouse model and the M of serum. Also, in their comparison of strain of EMC virus (6, 7). In mice, this diabetics with controls, low complement- strain of virus caused necrosis of the /3-cells fixation titers (>1:5) to Coxsackie B vi- of the islets, release of insulin, and, in some ruses were considered positive. Our experi- cases, a "diabetic-like state" characterized ence with this test suggests that many sera by hyperglycemia. Similar findings have at the 1:5 dilution fix complement with recently been reported by other investiga- crude Coxsackie B antigens nonspecifi- tors (22, 23). Genetic factors also seem to cally; hence, 1:5 "titers" might not be gen- be determinants in this model of "virus- uine indicators of past infection. induced" diabetes; only certain strains of In a separate study, Gamble and Taylor mice, when infected, will develop the le- showed a statistical correlation between sions suggestive of diabetes (24). the annual number of isolates of Coxsackie The fact that EMC virus is not a com- B-4 virus and the number of new cases of mon human pathogen appears to preclude juvenile diabetes (21). The correlation was the possibility that EMC virus itself is not seen for the other Coxsackie B viruses. involved in human diabetes. One study However, attempts to show a temporal conducted by Craighead et al. showed a 20 correlation between diabetes and other per cent incidence of EMC antibody in a viral infections common during the same Panamanian population (25), and a later season were not made. Thus this temporal survey of a random population of hospital- relationship remains open to question. ized children in the New England states revealed a 15 per cent incidence of EMC If diabetes mellitus is related etiologi- antibody (26). Indeed, in the present cally to a specific viral illness, one might study, only 10 per cent of the persons expect to see cases clustered in time follow- tested had measurable levels of EMC anti- ing an outbreak of that viral illness in a body (and none of these were diabetic population. One such study following a children). Coxsackie B-3 and B-4 mixed outbreak of 285 "OUTBREAK" OF JUVENILE DIABETES MELLITUS upper respiratory infection has recently volved, and no single exposure common to been completed; no cases of diabetes were the cases. Although seven diabetics had found among 109 children four years after family histories of diabetes, there was no the outbreak (30). Our purpose in the history of diabetes in their immediate present investigation was to determine families, and most of the relatives with that the 12 cases of diabetes in Pinellas diabetes apparently had adult onset dis- County represented an unusually high inci- ease. None of the diabetics had ex- dence and to look retrospectively for evi- perienced a recent viral illness, such as dence that a single viral agent may have measles or mumps, whose etiology could be been associated with the cluster. identified. Furthermore, none had ex- Previous reports suggesting a relation- perienced any abdominal symptoms char- ship between diabetes and viral infections acteristic of pancreatitis. Illnesses of pos- have dealt primarily with the juvenile form sible viral etiology were reported by eight of of the disease. We, likewise, confined our the diabetics. The illnesses, in general, investigation to cases of juvenile diabetes, were similar to common winter respiratory specifically in the pediatric population. illnesses. But because these infections For a number of reasons, we feel that our preceded symptoms of diabetes in only four evidence supporting an unusual incidence cases, we could not relate these illnesses to of juvenile diabetes in early 1972 is valid. the onset of diabetes. Firstly, our definition of a new case of The lack of viral isolates from the dia- juvenile diabetes was quite strict, and we betic children and their family members included in our analysis only new cases in was not unexpected; samples were taken which there was little doubt of the diagno- weeks after the recent "viral-like" illnesses sis, cases which in a modern medical com- and after the onsets of diabetes, at a time munity would be hospitalized. Secondly, when all individuals tested had recovered because the county where the investigation from their respiratory illnesses. It was took place was somewhat isolated geo- therefore necessary to rely upon serologic graphically and excellent medical facilities testing of single "convalescent" serum were available within the county, we feel it specimens in a search for evidence of re- unlikely that cases would have been hospi- cent viral infection. Results on sera from talized elsewhere. Thirdly, our case finding diabetic children revealed no elevated viral techniques, viz., review of medical records, antibody titers common to all. High titers should have been equally sensitive in individual patients to a number of throughout the 51-month period investi- different viruses were observed: influenza gated. Finally, rather than comparing the A, parainfluenza 3, respiratory syncytial incidence during early 1972 statistically virus, Coxsackie B-3, Coxsackie B-4, para- with the "expected" incidence over the influenza- 4B, cytomegalovirus, and coro- 4-year period, we compared it with the navirus 229-E. Eight of the nine diabetics previous maximum incidence. Since the had at least one high titer suggestive of a maximum occurred in 1971 and the Pinel- recent viral infection. Whether these had las County population has been stable, been clinical infections or inapparent in- growth in the pediatric population could fections could not be determined. There not likely explain the peak of cases ob- was no evidence of recent mumps or EMC served in 1972. infection in any of the diabetics. The detailed investigation of nine of the In comparison with non-diabetic con- 12 cases which comprised the "outbreak" trols, matched closely for time of serum in early 1972 revealed no geographic clus- collection, residence, age and sex, antibody tering of cases within the county, no per- to only one virus appeared more prevalent sonal association between the families in- in diabetics. Three of nine diabetics had 286 HUFF, HIERHOLZER, FARRIS Diabetic ketoacidosis associated with mumps high titers to Coxsackie B-3. Although the virus infection. Ann Intern Med 78:663-667, 1973 prevalence of high titers was significantly 4. Forrest JM, Menser MA, Harley JD: Diabetes greater in the diabetics, the GMT of the mellitus and congenital rubella. Pediatrics group was not significantly different from 44:445-446, 1969 matched controls. Family members of dia- 5. Johnson GM, Tudor RB: Diabetes mellitus and congenital rubella infection. Am J Dis Child betics possessed high titers to a related 120:453-455, 1970 virus, Coxsackie B-4. Since antibody to 6. Craighead JE, McLane MF: Diabetes mellitus: some of the Coxsackie B viruses can remain induction in mice by encephalomyocarditis virus. elevated for years following infection, the Science 162:913-914, 1968 significance of the elevated B-3 and B-4 SN 7. Craighead JE, Steinke J: Diabetes mellitus-like syndrome in mice infected with enceph- titers observed in our study is much more alomyocarditis virus. Am J Pathol 63:119-134, questionable than antibody levels to the other viruses. Perhaps Coxsackie B infec- 8. Gamble DR, Kinsley ML, Fitzgerald MG, et al: tions were related to some of the cases of Viral antibodies in diabetes mellitus. Br Med J diabetes, but we have no evidence that 3:627-630, 1969 9. Hierholzer JC, Mostow SR, Dowdle WR: Prospec- Coxsackie B viruses were responsible for tive study of a mixed coxsackie virus B3 and B4 the "outbreak" of diabetes. Furthermore, outbreak of upper respiratory illness in a chil- the cases of diabetes in this investigation dren's home. Pediatrics 49:744-752, 1972 occurred six months after the usual late 10. Robinson RQ, Dowdle WR: Influenza viruses. In summer-early fall peak of Coxsackie B Diagnostic Procedures for Viral and Rickettsial Infections. Edited by EH Lennette and NJ infections (9). Schmidt. New York, American Public Health Temporally, the cluster of cases of diabe- Association, 1969, p 414-433 tes followed closely a community-wide in- 11. Casey HL: Standardized diagnostic complement fluenza outbreak and the usual season of fixation method and adaptation to micro test. winter respiratory viral infections. A clus- Public Health Monograph No. 74, U.S. Public Health Service, Washington, D.C., 1965 ter of cases of diabetic ketoacidosis during 12. Hierholzer JC, Suggs MT: Standardized viral an influenza epidemic in the winter of hemagglutination and hemagglutination-inhibi- 1969-1970 has been previously reported tion tests. I. Standardization of erythrocyte sus- (31). Six of 29 cases described in this report pensions. Appl Microbiol 18:816-823, 1969 (31) were in individuals previously un- 13. Hierholzer JC, Suggs MT, Hall EC: Standardized viral hemagglutination and hemagglutination- known to be diabetics. Despite the tem- inhibition tests. II. Description and statistical poral association in our study, we were able evaluation. Appl Microbiol 18:824-833, 1969 to demonstrate an elevated titer to influ- 14. Killgore GE, Dowdle WR: Antigenic characteriza- enza in only one diabetic and in only three tion of parainfluenza 4A and 4B by the hemag- of the family members. glutination-inhibition test and distribution of HI antibody in human sera. Am J Epidemiol Our results, therefore, do not implicate 91:308-316, 1970 any particular virus in this "outbreak" of 15. Bernstein MT, Stewart JA: Method for typing diabetes. If viruses are indeed etiologically antisera to herpesvirus hominis by indirect he- involved, the data at most suggest an effect magglutination-inhibition. Appl Microbiol of multiple viruses rather than any single 21:680-684, 1971 16. Bernstein MT, Stewart JA: Indirect hemaggluti- entity upon genetically predisposed indi- nation test for detection of antibodies to cyto- viduals. megalovirus. Appl Microbiol 21:84-89, 1971 17. Kaye HS, Ong SB, Dowdle WR: Detection of REFERENCES coronavirus 229E antibody by indirect hemag- 1. Hinden E: Mumps followed by diabetes. Lancet glutination. Appl Microbiol 24:703-707, 1972 1:1381, 1962 18. Steinberg AG, Rushforth NB, Bennett PH: On 2. McCrae WM: Diabetes mellitus following genetics of diabetes mellitus. In Pathogenesis of mumps. Lancet 1:1300-1301, 1963 Diabetes Mellitus. Edited by E Cerasi and R Luft. Stockholm, Almquist and Wiksell, 1970 3. Block MB, Berk JE, Fridhandler LS, et al: "OUTBREAK" OF JUVENILE DIABETES MELLITUS 287 19. Pathogenesis of diabetes mellitus. Br Med J Med 114:500-503, 1963 3:594-595, 1971 26. Craighead, JE : personal communication, 1974 20. Adams SF: The seasonal variation in the onset of 27. Pappenheimer AM, Kunz, LJ, Richardson S: acute diabetes. Arch Intern Med 37:861-864, 1926 Passage of coisackie virus (Conn-5 strain) in 21. Gamble DR, Taylor KW: Seasonal incidence of adult mice with production of pancreatic dis- diabetes mellitus. Br Med J 3:631-633, 1969 ease. J Exp Med 94:45-65, 1951 22. Muntefering H: Zur pathologie des diabetes mel- 28. Coleman TJ, Gamble DR, Taylor KW: Diabetes litus der wei Ben Maus bei der EMC-Virusinfek- in mice after coxsackie B4 virus infection. Br Med tion. Virchows Arch Abt A Pathol Anat J 3:25-27, 1973 356:207-234, 1972 29. Kibrick S, Benirschke K: Severe generalized 23. Boucher DW, Notkins AL: Virus-induced diabe- disease (encephalohepatomyocarditis) occurring tes mellitus. I. Hyperglycemia and hypoin- in the newborn period and due to infection with sulinemia in mice infected with enceph- coxsackie virus, group B. Pediatrics 22:857-875, alomyocarditis virus. J Exp Med 137:1226-1239, 1958 1973 30. Hierholzer JC, Farris WA: Follow-up of children 24. Craighead JE: Workshop on viral infection and infected in a coxsackievirus B-3 and B-4 out- diabetes mellitus in man. J Infect Dis break: no evidence of diabetes mellitus. J Infect 125:568-570, 1972 Dis 129:741-746,1974 25. Craighead JE, Peralta PH, Shelokov A: Demon- 31. Watkins PJ, Soler NG, Fitzgerald MG, et al: stration of encephalomycarditis virus antibody in Diabetic ketoacidosis during the influenza epi- human serums from Panama. Proc Soc Exp Biol demic. Br Med J 4:89-91, 1970 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Pubmed Central

AN “OUTBREAK” OF JUVENILE DIABETES MELLITUS: CONSIDERATION OF A VIRAL ETIOLOGY

HUFF, J. CLARK; HIERHOLZER, JOHN C.; FARRIS, WILLIAM A.
American Journal of Epidemiology , Volume 100 (4) – Oct 1, 1974
Download PDF
11 pages

Loading next page...
 
/lp/pubmed-central/an-outbreak-of-juvenile-diabetes-mellitus-consideration-of-a-viral-7IG5kRNeMw

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Pubmed Central
Copyright
© 1975 by The School of Hygiene and Public Health of The Johns Hopkins University
ISSN
0002-9262
eISSN
1476-6256
DOI
10.1093/oxfordjournals.aje.a112036
Publisher site
See Article on Publisher Site

Abstract

AMERICAN JOURNAL OF EPIDEMIOLOGY Vol. 100, No. 4 Copyright © 1974 by The Johns Hopkins University Printed in U.SA. AN "OUTBREAK" OF JUVENILE DIABETES MELLITUS: CONSIDERATION OF A VIRAL ETIOLOGY J. CLARK HUFF, JOHN C. HLERHOLZER AND WILLIAM A. FARRIS Huff. J . C , J. C. Hierholzer (Respiratory Virology Section, Center for Disease Control. Atlanta, Ga. 30333) and W. A. Farris. An "outbreak" of juvenile diabetes mellitus: consideration of a viral etiology. Am J Epidemiol 100:277-287 . 1974.—Nine of twelve cases of juvenile diabetes mellitus, representing an unusual geographic and temporal cluster, were investigated for evidence that a specific viral infection might be etiologically related to their occurrence. Eight diabetics had experienced recent "viral-like" illnesses, predominantly repiratory, but these illnesses bore no uniform temporal relation to their onsets of diabetes. Diabetics demonstrated no serologic evidence of a recent viral illness common to all. Elevated titers to only one virus, coxsackie B-3, were more prevalent in diabetics than in controls (33% vs. 6%), but geometric mean titers of diabetics to a panel of 26 common viral antigens were similar to those of controls. These data neither support nor negate the hypothesis that infection with a specific virus precipitates juvenile diabetes mellitus. diabetes mellitus, juvenile; respiratory tract infections INTRODUCTION co-workers (6, 7) have provided experimen- tal evidence that mice infected with a Recent experimental and epidemiologic strain of encephalomyocarditis (EMC) data have generated a growing interest in virus may develop damage of the /9-cells of the possibility that viral infections may be the pancreatic islets and a resultant "dia- involved in the pathogenesis of diabetes betic-like" state. Furthermore, Gamble mellitus. Scattered case reports have sug- and co-workers (8) have reported that the gested that human diabetes may follow Coxsackie B viruses, picornaviruses similar such viral diseases as mumps (1-3) and to EMC virus, may be associated with congenital rubella (4, 5). Craighead and human diabetes. Gamble's data included higher prevalence of antibody to Coxsackie Received for publication February 26, 1974, and in B viruses in individuals with juvenile dia- final form May 28, 1974. 1 From the Viral Diseases Branch, Bureau of betes of recent onset than in controls or Epidemiology and the Virology Branch, Bureau of long-term diabetics. Laboratories, Center for Disease Control, Atlanta, Ga. In early 1972, a group of pediatricians in Abbreviations: CF, complement fixation; CMV, Pinellas County, Florida, noted a large cytomegalovirus; EMC, encephalomyocarditis; GMT, number of new cases of juvenile diabetes. geometric mean titer; HI, hemagglutination-inhibi- This report describes the results of an tion; IHA, indirect hemagglutination; MKC, monkey kidney cell; RSV, respiratory syncytial virus; SN, epidemiologic and laboratory investigation serum neutralization. of these cases. The authors gratefully acknowledge the support of Dr. J. P. Hanley, a pediatrician in Clearwater, Flor- BACKGROUND ida; Dr. G. M. Dame, Director, Pinellas County Health Department; and other physicians in Pinellas Pinellas County lies along the Gulf Coast County; and gratefully acknowledge the technical of Florida just north of the mouth of assistance of J. T. Cole, K. Hilliard, and W. B. Yarbrough. Tampa Bay. The flat lowlands of the 277 278 HUFF, HIERHOLZER, FARRIS county are covered by contiguous urban positive viral serologic tests from Pinellas and suburban areas, where 96 per cent of County were also reviewed for the same the county's half million residents live period. (1970 census). The largest cities are St. Laboratory. In mid-March 1972, rectal Petersburg (population 216,232) and Clear- and throat swabs were taken from newly water (population 52,074). Tourism is the diabetic children and other members of main industry in the county. Socioeconom- their households. Specimens placed in ically, most residents are middle class. tryptose phosphate broth with 0.5 per cent Because the county is situated on a penin- gelatin were frozen in Dry Ice. In the sula and excellent medical care is available laboratory, specimens were processed for in the county, Pinellas County residents inoculation onto two strains of human generally seek medical care within Pinellas diploid lung fibroblast (RU-1 and WI-38), County. one line of human epithelium (HEp-2), and primary rhesus monkey kidney cell METHODS (MKC) cultures, as previously described (9). The specimens were also inoculated Epidemiologic. Medical records at the into 10-day embryonated eggs (10). Each seven hospitals in Pinellas County where specimen was passaged at least three times children are usually admitted were re- for a total incubation period of four weeks. viewed. Records of individuals up to age 14 who were admitted between January 1968 Single (convalescent) serum specimens and March 1972 as cases of diabetes mel- were collected by venipuncture in mid- litus were examined. A new "case" of March from newly diabetic children, mem- juvenile diabetes was defined as an illness bers of their households, and a control 1) diagnosed as diabetes in a child not group of children from Pinellas County previously known to have the disease, 2) with no history of diabetes. Non-diabetic characterized by glycosuria and hyper- controls were matched with the newly glycemia (a blood glucose of a least 200 diabetic children for age and sex. Serum mg/100 ml), and 3) treated with insulin at samples were stored at -30 C until tested. the time of hospital discharge. The "inci- The sera were tested for antibodies to the dence" of juvenile diabetes mellitus in the prototype strains of a large number of pediatric age group was estimated by plot- respiratory viruses by complement-fixation ting all new cases by month of hospitaliza- (CF), hemagglutination-inhibition (HI), tion. No effort was made to identify new indirect hemagglutination (IHA), or serum cases in older age groups or new cases in neutralization (SN) tests, as appropriate persons not hospitalized. for each particular virus. The sera were Investigation of cases. Diabetic children heat-inactivated (56 C, 30 minutes) for all hospitalized early in 1972 were investigated serologic tests. Antibodies to influenza A in detail. With the consent of the child's and B, parainfluenza 1, 2, and 3, the parents, a detailed medical history was "soluble" and "viral" antigens of mumps taken from each family member. Each virus, the group-specific hexon antigen of family was questioned concerning associa- adenovirus, respiratory syncytial virus tions between them and the families of (RSV), Mycoplasma pneumoniae, herpes other diabetic children. simplex 1, and cytomegalovirus (CMV) Review of public health records. Physi- were assayed by the standardized CF test cians' reports of viral diseases to the county with overnight fixation of 5 units of com- health department were examined for the plement (11). Antibodies to influenza C, time period November 1971 through March parainfluenza 4A and 4B, and coronavirus 1972. Records of positive viral cultures and OC 43 were assayed by the standardized HI "OUTBREAK" OF JUVENILE DIABETES MELLITUS test with .01 M phosphate-buffered saline case every 2 months. The maximum num- diluent and spectrophotometrically stand- ber in a consecutive 2-month period was ardized 0.4 per cent mammalian or 0.5 per four in 1971. In February and March 1972, cent avian red blood cells (12, 13). Hemag- 12 cases occurred. Based on an "expected" glutinating antigens of parainfluenza 4A rate of four new cases or fewer in a 2-month and 4B were prepared for use in th e HI test period, the occurrence of 12 cases was a as described by Killgore and Dowdle (14). very rare event (p < .001). Antibodies to herpes simplex 1 and CMV Investigation of cases. Nine of the 12 were determined by th e IHA procedures of patients hospitalized in February and Bernstein and Stewart (15, 16) and corona- March and their families agreed to partici- virus 229E antibodies were measured by a pate in the detailed case investigations. similar IHA test (17). Neutralizing anti- The nine children with newly diagnosed bodies to the six Coxsackie B viruses were diabetes included five males and four fe- assayed by SN tests in MKC tissue culture males, ages 4 to 13 (mean age 8.3). Al- (9). Neutralizing antibody titers to EMC though all nine had been hospitalized at virus (r+ variant), assayed by plaque neu- either of two hospitals in central Pinellas tralization tests in L-cell tissue culture, County, their places of residence were were generously provided by Dr. John E. scattered throughout the county. No dia- Craighead, University of Vermont. betic child and family was acquainted with any other child and family, and none were RESULTS associated in educational, religious, occu- Epidemiologic. Thirty-five cases of juve- pational, or recreational activities. No nile diabetes diagnosed in children be- medication or toxic exposure was common tween January 1968 and March 1972 are to the nine. presented by month of hospital admission A summary of data from case investiga- in figure 1. In the 48 months prior to tions is shown in table 1. Onset of diabetic January 1972, 23 new diabetics were hospi- symptoms preceded hospitalization by 1 to talized, an average "incidence" of one new 7 weeks (mean 3 weeks). Two factors com- 1O 8 - in in 3 6- 4 - 2 - JFMAMJ JASONDlJFMAMJ J ASONDlj FMAM J JASOND JFMAMJ J AS ONDljFM 1968 1969 1970 1971 1972 Month of Hospitalization FIGURE 1. Thirty-five cases of newly diagnosed juvenile diabetes mellitus, by month of hospital admission, Pinellas County, Florida, January 1, 1968-March 31, 1972. 280 HUFF, HIERHOLZER, FARRIS TABLE 1 Summary of epidemiologic data for nine new cases of juvenile diabetes Viral illness in 3 Date of months preceding hos- initial Date of pitalization h capital- Family history of symp- Patient Age Sex ization diabetes Date Type toms* 1 6 M 1/16* 2/1 Maternal great-aunt 12/29 URI.t otitis media 2 4 F 1/29 2/11 Paternal aunt, paternal 12/25 "influenza-like" great-grandmother illness 3 11 2/11 2/7 F 1/11 Maternal grandfather Pharyngitis 4 9 F 2/7 2/14 Maternal and paternal 1/15 Gastroenteritis cousins 6 M 2/16 Paternal grandmother 12/25 URI 5 12/25 6 10 M 2/8 2/22 None 2/18 URI 7 12 2/14 3/4 Maternal grandmother 3/4 ? URI, otitis media 8 13 F 2/18 3/9 Unknown (adopted) None None 4 M 2/24 3/9 Maternal grandmother 2/15 "influenza-like" illness * December 1971 to March 1972. t URI, upper respiratory infection. mon to the majority of the diabetics were a in Clearwater in December 1971, and influ- family history of diabetes and an illness in enza was recognized throughout the county the 3 months preceding hospitalization in January 1972. Outbreaks of other report- considered by the patient's physician to be able viral illnesses, including mumps and of possible viral etiology. Seven of the nine rubella, had not occurred. Influenza was families knew of relatives with diabetes, confirmed by isolation of A/Hong Kong/68 but further information concerning the (H3N2) strains and by positive serologic diabetic relatives was unavailable. Two of studies. Positive serologic tests were also the diabetic relatives were reported to have obtained for parainfluenza and adenovirus. had onset of disease in young adulthood; ECHO-11 and Coxsackie A-12 were iso- the others apparently were adult onset, lated from individual patients, and in ad- insulin-independent diabetics. dition, Coxsackie B-2 and ECHO-8 were Eight of the diabetics had histories of a isolated from sewage effluents. With the preceding illness of possible viral etiology. exception of influenza, however, norre of The etiology of the illnesses was not deter- these agents was known to be associated mined. Seven of the illnesses were respira- with significant morbidity. tory, ranging from simple upper respiratory Laboratory. Cultures of rectal and throat to febrile lower respiratory illnesses. In swabs taken from eight newly diabetic only four of the eight cases, however, did children and nine of their family members the illness precede the onset of diabetic yielded no virus isolates. At the time of symptoms. attempted isolation in mid-March, there Review of public health records. In a was no respiratory illness in any of the search for viral illnesses prevalent in Pinel- individuals tested; isolation attempts were las County prior to the cluster of cases of therefore made only on the chance that juvenile diabetes, reports to the county there might still be virus excretion from a health department of viral diseases were preceding illness. reviewed. An outbreak of measles occurred Single serum samples were taken from "OUTBREAK" OF JUVENILE DIABETES MELLTTUS 281 nine diabetic patients within 6 weeks after high titers of antibody to 15 viruses and to hospitalization and within 12 weeks after M. pneumoniae were quite similar in dia- onset of symptoms. Unfortunately, earlier betics and controls for all agents except sera for either the respiratory or diabetes Coxsackie B-3 (table 3). Three of nine illnesses were not available. Single serum diabetics (33 per cent) had significant specimens were also collected from 21 titers (> 1:80) to this virus, whereas only 2 family members (contacts) and from 36 of 36 matched, non-diabetic controls (6 per matched and 20 unmatched control chil- cent) had similar titers. This difference dren at this time (table 2). All of these sera were tested for antibody to 15 common TABLE 2 respiratory viruses and to M. pneumoniae. Age and sex characteristics of study population The antibody levels of the case, contact, Sex Age and matched control groups were tabu- Fe- Me- lated both as numbers of persons with Total Male male dian Range "significantly high" antibody titers (table Diabetics 9 5 4 9 3) and as geometric mean titers (GMT) 4-13 Family con- 21 10 26 3-46 (table 4). tacts "Significant" antibody titers were de- Controls, 36 20 16 9 ' 4-13* fined as those levels of antibody (serum matched Controls, 56 8 endpoint dilutions) which we considered, 23 33 4-13 total through diagnostic experience with the test procedures used, to be suggestive of recent * Diabetics and controls matched for age (within infection. The prevalence of significantly 1 year), sex, locality, and time of serum collection. TABLE 3 Prevalence of elevated antibody titers in diabetic patients, family members, and matched controls % with elevated titer (significant) 21 family 36 matched Virus Test titer" 9 patients members controlst Influenza A CF 1:32 11 14 19 Influenza B CF 1:32 0 0 Parainfluenza 1 CF 1:32 0 3 Parainfluenza 2 CF 1:32 0 3 Parainfluenza 3 CF 1:32 11 0 8 Mumps (soluble) CF 1:16 0 5 5 Mumps (viral) CF 1:32 0 0 Adenovirus CF 1:16 0 14 6 RSV CF 1:32 11 14 14 Herpes 1 CF 1:32 0 10 7 0 0 16 M. pneumoniae CF 1:32 0 10 4 Coxsackie B-l SN 1:80 Coxsackie B-2 SN 1:80 0 14 13 Coxsackie B-3 SN 33 6 1:80 10 Coxsackie B-4 SN 11 17 1:80" 38 Coxsackie B-5 SN 0 4 1:80 0 Coxsackie B-6 SN 1:80 0 0 * A titer of this level or higher is considered elevated and therefore may be an indication of recent infection (within approximately 5 months). See text for complete definition. t The frequency of elevated titers in the diabetic and control groups was not significantly different at the .05 level by Fisher's Exact Test for any virus except Coxsackie B-3 (P = .047). 282 HUFF, HIERHOLZER, FARRIS was of borderline statistical significance (p Results of GMT's to nine additional viral = .047) by Fisher's Exact Test. antigens for all diabetics, selected family members, and an unmatched group of Family members of the diabetics were a non -diabetic children are shown in table 5. diverse group of 12 adults and nine chil- As in table 4, antibody levels to most dren and therefore were not compared with viruses were typical of pediatric groups. the control group except in general terms. The GMT to CMV, which was somewhat Elevated titers to adenovirus, herpesvirus, high in diabetics, was nonetheless quite and Coxsackie virus B-4 were most preva- similar to that of the non-diabetic group lent among family members. and was not in an unusual range. The CF The serologic data expressed as GMT's and IHA tests for herpes and for CMV to 17 viral antigens did not reveal any measure different aspects of infection, and single agent to which diabetics possessed a neither test indicated that these viruses significantly higher mean titer (table 4). were associated with the respiratory ill- The GMT of diabetics to Coxsackie B-3 nesses of the diabetics before they were was almost two dilutions higher than con- hospitalized. trols, but statistically, the difference was not significant by the t-test (p = .083). The Antibody levels to the six Coxsackie B GMT's of family members, despite their viruses are presented for each of the nine age and sex diversity, were similar to those cases, together with the serologic identifi- of diabetics and controls for all agents ex- cation of other possible recent respiratory cept Coxsackie B-4; the GMT of family illnesses, to provide a summary of the most members to Coxsackie B-4 was approxi- salient serologic data (table 6). The pattern mately one dilution greater than the GMT seen in this summary is one of total ran- of diabetics and controls. domness. Virtually each case had serologic TABLE 4 Geometric mean serum antibody titers to 17 viral antigens in diabetic patients, family members, and controls GMT* 21 family 36controlB 56 controls Virus Test 9 patients members (matched only)f (total) Influenza A CF 6.86 8.84 6.92 7.38 CF Influenza B 4.00 4.56 4.16 4.16 CF 4.72 Paiainfluenz a 1 4.32 4.58 4.88 CF 5.21 5.32 Parainfluenz a 2 5.44 5.14 Parainfluenz a 3 CF 8.64 5.75 8.00 7.90 Mump s (soluble) CF 5.88 4.93 5.83 5.80 Mump s (viral) CF 5.88 4.56 8.32 8.31 Adenovirus CF 5.04 5.94 5.99 6.09 RSV CF 12.70 9.76 9.33 9.63 Herpes 1 CF 4.00 7.74 5.24 5.52 M . pneumoniae CF 4.67 5.75 8.64 8.83 Coxsackie B-l 5.00 7.70 5.52 5.56 SN Coxsackie B-2 SN 6.15 8.94 13.53 13.75 Coxsackie B-3 SN 23.04 9.61 8.70 8.92 Coxsackie B-4 SN 25.05 10.40 11.56 14.70 Coxsackie B-5 SN 5.63 5.79 5.74 6.30 SN 5.61 5.44 5.14 Coxsackie B-6 5.00 * A geometric mean titer (GMT) of 4.00 = < 1:8 for CF tests and of 5.00 - < 1:10 for SN tests. GMT's are listed as dilution factor of endpoint serum dilutions. t The GM T in th e diabetic and matched control groups was no t significantly different (P > .05) for any virus by the (-test. "OUTBREAK" OF JUVENILE DIABETES MELUTUS TABLE 5 Prevalence of "significant" antibody titers in diabetics and geometric mean serologic titers of diabetics, selected family members, and selected unmatched controls to 9 viral antigens Geometric mean titers* Prevalence of significant 10 family 24 Significant titers in 9 9 diabetics members controls Virus Test titer diabetics (%) (ages 4-13) (ages 5-23) (ages 4-13) Influenza C HI 1:128 0 13.72 19.70 NDf Parainfluenza 4A HI 1:128 0 13.39 13.00 ND Parainfluenza 4B HI 1:128 11 9.69 12.96 ND Herpes 1 IHA 1:2048 0 6.35 14.04 6.50 Cytomegalovirus CF 1:256 22 46.31 28.89 22.42 IHA 1:2048 11 21.82 12.96 26.90 Coronavirus OC43 HI 1:64 ND 0 11.91 14.60 Coronavirus 229E IHA 1:64 11 ND 9.83 8.58 Encephalomyocarditis PN§ 1:4 1.31| 0 l.OOt 1.351: • GMT of 4.00 - <1:8. GMT's are listed as dilution factor of endpoint serum dilution. t Not done. t GMT of 1.00 - <1:2. 5 PN - plaque neutralization. TABLE 6 Summary of Coxsackie B antibody titers and significant titers to other viruses for 9 new cases of juvenile diabetes mellitus SN antibody titers to Coisaclde virus* Case B-l B-2 B-3t B-5 B-6 viruses B-4t 1 0 0 160 0 0 None 0* 2 0 10 0 40 0 0 CMV 3 0 0 20 160 0 0 None 4 0 0 0 0 10 0 RSV, parainfluenza 4B 5 0 0 40 0 0 0 Coronavirus 229E 6 0 0 160 40 0 0 CMV 7 0 20 80 0 0 0 Influenza A 8 0 40 0 20 20 0 Parainfluenza 3 0 0 9 0 0 0 0 None * Serum neutralization (SN) titers listed as dilution factor of endpoint serum dilutions, t Among the 56 controls, 37 had SN titers to B-3 of < 1:10, 9 had titers of 1:10-1:20, 7 had 1:40-1:80, and 3 were >l:160; 35 had SN titers to B-4 of <l:10, 5 had 1:10-1:20, 12 had 1:40-1:80, and 4 were >l:160. evidence of infection with a different virus has not yet been clarified (18). Some inves- in the preceding four to six months. It is tigators have concluded that both genetic therefore apparent that no one of the and environmental determinants are in- Coxsackie B viruses nor any one of the 23 volved (19). Among the characteristics viruses tested for could account for this suggestive of environmental determinants "outbreak" of diabetes. is a seasonality in the onset of juvenile diabetes. Both a 1926 study of patients from the north central United States (20) DISCUSSION and a 1969 analysis of English diabetics Heredity has been accepted as the major (21) described a nadir in the incidence in determinant of the occurrence of diabetes late spring and early summer and a broad mellitus, but the exact mode of inheritance 284 HUFF, HIERHOLZER, FARRIS peak between late summer and early Attention has been focused therefore spring. Data collected in Pinellas County upon a similar group of picornaviruses, the also hinted of a seasonal pattern: 11 of the Coxsackie B viruses, which are extremely 23 cases identified from 1968 through 1971 frequent human pathogens. The Coxsackie (figure 1) were hospitalized in the months B viruses, like EMC, may cause pancreati- January-April, approximately the time of tis (27) and even diabetes (28) in experi- the 1972 peak. Many viral respiratory in- mental mice, and, in addition, they may fections have a similar seasonal pattern, attack the human pancreas in certain in- and that characteristic has suggested a fant infections (29). Although there are no possible cause of the seasonality of juvenile published case reports of diabetes in hu- diabetes. mans following proven Coxsackie B infec- tions, the epidemiologic and serologic data Although case reports have indicated of Gamble and co-workers (8) aroused that a diabetic state may follow such considerable interest in a possible relation diseases as mumps and congenital rubella between these viruses and human diabetes. (1-5), these reports have remained inciden- Gamble et al. reported that insulin- tal observations. There is, at present, no dependent diabetics within 3 months of firm evidence that one specific viral illness onset had higher serum neutralization ti- plays a significant part in the pathogenesis ters to Coxsackie B viruses, especially B-4, of human diabetes mellitus. than did non-diabetic controls or long- The most convincing data suggesting a standing diabetics (8). Controls in this viral etiology of diabetes come from the study, however, were not closely matched experiments carried out by Craighead and for age, sex, residence, or time of collection co-workers using a mouse model and the M of serum. Also, in their comparison of strain of EMC virus (6, 7). In mice, this diabetics with controls, low complement- strain of virus caused necrosis of the /3-cells fixation titers (>1:5) to Coxsackie B vi- of the islets, release of insulin, and, in some ruses were considered positive. Our experi- cases, a "diabetic-like state" characterized ence with this test suggests that many sera by hyperglycemia. Similar findings have at the 1:5 dilution fix complement with recently been reported by other investiga- crude Coxsackie B antigens nonspecifi- tors (22, 23). Genetic factors also seem to cally; hence, 1:5 "titers" might not be gen- be determinants in this model of "virus- uine indicators of past infection. induced" diabetes; only certain strains of In a separate study, Gamble and Taylor mice, when infected, will develop the le- showed a statistical correlation between sions suggestive of diabetes (24). the annual number of isolates of Coxsackie The fact that EMC virus is not a com- B-4 virus and the number of new cases of mon human pathogen appears to preclude juvenile diabetes (21). The correlation was the possibility that EMC virus itself is not seen for the other Coxsackie B viruses. involved in human diabetes. One study However, attempts to show a temporal conducted by Craighead et al. showed a 20 correlation between diabetes and other per cent incidence of EMC antibody in a viral infections common during the same Panamanian population (25), and a later season were not made. Thus this temporal survey of a random population of hospital- relationship remains open to question. ized children in the New England states revealed a 15 per cent incidence of EMC If diabetes mellitus is related etiologi- antibody (26). Indeed, in the present cally to a specific viral illness, one might study, only 10 per cent of the persons expect to see cases clustered in time follow- tested had measurable levels of EMC anti- ing an outbreak of that viral illness in a body (and none of these were diabetic population. One such study following a children). Coxsackie B-3 and B-4 mixed outbreak of 285 "OUTBREAK" OF JUVENILE DIABETES MELLITUS upper respiratory infection has recently volved, and no single exposure common to been completed; no cases of diabetes were the cases. Although seven diabetics had found among 109 children four years after family histories of diabetes, there was no the outbreak (30). Our purpose in the history of diabetes in their immediate present investigation was to determine families, and most of the relatives with that the 12 cases of diabetes in Pinellas diabetes apparently had adult onset dis- County represented an unusually high inci- ease. None of the diabetics had ex- dence and to look retrospectively for evi- perienced a recent viral illness, such as dence that a single viral agent may have measles or mumps, whose etiology could be been associated with the cluster. identified. Furthermore, none had ex- Previous reports suggesting a relation- perienced any abdominal symptoms char- ship between diabetes and viral infections acteristic of pancreatitis. Illnesses of pos- have dealt primarily with the juvenile form sible viral etiology were reported by eight of of the disease. We, likewise, confined our the diabetics. The illnesses, in general, investigation to cases of juvenile diabetes, were similar to common winter respiratory specifically in the pediatric population. illnesses. But because these infections For a number of reasons, we feel that our preceded symptoms of diabetes in only four evidence supporting an unusual incidence cases, we could not relate these illnesses to of juvenile diabetes in early 1972 is valid. the onset of diabetes. Firstly, our definition of a new case of The lack of viral isolates from the dia- juvenile diabetes was quite strict, and we betic children and their family members included in our analysis only new cases in was not unexpected; samples were taken which there was little doubt of the diagno- weeks after the recent "viral-like" illnesses sis, cases which in a modern medical com- and after the onsets of diabetes, at a time munity would be hospitalized. Secondly, when all individuals tested had recovered because the county where the investigation from their respiratory illnesses. It was took place was somewhat isolated geo- therefore necessary to rely upon serologic graphically and excellent medical facilities testing of single "convalescent" serum were available within the county, we feel it specimens in a search for evidence of re- unlikely that cases would have been hospi- cent viral infection. Results on sera from talized elsewhere. Thirdly, our case finding diabetic children revealed no elevated viral techniques, viz., review of medical records, antibody titers common to all. High titers should have been equally sensitive in individual patients to a number of throughout the 51-month period investi- different viruses were observed: influenza gated. Finally, rather than comparing the A, parainfluenza 3, respiratory syncytial incidence during early 1972 statistically virus, Coxsackie B-3, Coxsackie B-4, para- with the "expected" incidence over the influenza- 4B, cytomegalovirus, and coro- 4-year period, we compared it with the navirus 229-E. Eight of the nine diabetics previous maximum incidence. Since the had at least one high titer suggestive of a maximum occurred in 1971 and the Pinel- recent viral infection. Whether these had las County population has been stable, been clinical infections or inapparent in- growth in the pediatric population could fections could not be determined. There not likely explain the peak of cases ob- was no evidence of recent mumps or EMC served in 1972. infection in any of the diabetics. The detailed investigation of nine of the In comparison with non-diabetic con- 12 cases which comprised the "outbreak" trols, matched closely for time of serum in early 1972 revealed no geographic clus- collection, residence, age and sex, antibody tering of cases within the county, no per- to only one virus appeared more prevalent sonal association between the families in- in diabetics. Three of nine diabetics had 286 HUFF, HIERHOLZER, FARRIS Diabetic ketoacidosis associated with mumps high titers to Coxsackie B-3. Although the virus infection. Ann Intern Med 78:663-667, 1973 prevalence of high titers was significantly 4. Forrest JM, Menser MA, Harley JD: Diabetes greater in the diabetics, the GMT of the mellitus and congenital rubella. Pediatrics group was not significantly different from 44:445-446, 1969 matched controls. Family members of dia- 5. Johnson GM, Tudor RB: Diabetes mellitus and congenital rubella infection. Am J Dis Child betics possessed high titers to a related 120:453-455, 1970 virus, Coxsackie B-4. Since antibody to 6. Craighead JE, McLane MF: Diabetes mellitus: some of the Coxsackie B viruses can remain induction in mice by encephalomyocarditis virus. elevated for years following infection, the Science 162:913-914, 1968 significance of the elevated B-3 and B-4 SN 7. Craighead JE, Steinke J: Diabetes mellitus-like syndrome in mice infected with enceph- titers observed in our study is much more alomyocarditis virus. Am J Pathol 63:119-134, questionable than antibody levels to the other viruses. Perhaps Coxsackie B infec- 8. Gamble DR, Kinsley ML, Fitzgerald MG, et al: tions were related to some of the cases of Viral antibodies in diabetes mellitus. Br Med J diabetes, but we have no evidence that 3:627-630, 1969 9. Hierholzer JC, Mostow SR, Dowdle WR: Prospec- Coxsackie B viruses were responsible for tive study of a mixed coxsackie virus B3 and B4 the "outbreak" of diabetes. Furthermore, outbreak of upper respiratory illness in a chil- the cases of diabetes in this investigation dren's home. Pediatrics 49:744-752, 1972 occurred six months after the usual late 10. Robinson RQ, Dowdle WR: Influenza viruses. In summer-early fall peak of Coxsackie B Diagnostic Procedures for Viral and Rickettsial Infections. Edited by EH Lennette and NJ infections (9). Schmidt. New York, American Public Health Temporally, the cluster of cases of diabe- Association, 1969, p 414-433 tes followed closely a community-wide in- 11. Casey HL: Standardized diagnostic complement fluenza outbreak and the usual season of fixation method and adaptation to micro test. winter respiratory viral infections. A clus- Public Health Monograph No. 74, U.S. Public Health Service, Washington, D.C., 1965 ter of cases of diabetic ketoacidosis during 12. Hierholzer JC, Suggs MT: Standardized viral an influenza epidemic in the winter of hemagglutination and hemagglutination-inhibi- 1969-1970 has been previously reported tion tests. I. Standardization of erythrocyte sus- (31). Six of 29 cases described in this report pensions. Appl Microbiol 18:816-823, 1969 (31) were in individuals previously un- 13. Hierholzer JC, Suggs MT, Hall EC: Standardized viral hemagglutination and hemagglutination- known to be diabetics. Despite the tem- inhibition tests. II. Description and statistical poral association in our study, we were able evaluation. Appl Microbiol 18:824-833, 1969 to demonstrate an elevated titer to influ- 14. Killgore GE, Dowdle WR: Antigenic characteriza- enza in only one diabetic and in only three tion of parainfluenza 4A and 4B by the hemag- of the family members. glutination-inhibition test and distribution of HI antibody in human sera. Am J Epidemiol Our results, therefore, do not implicate 91:308-316, 1970 any particular virus in this "outbreak" of 15. Bernstein MT, Stewart JA: Method for typing diabetes. If viruses are indeed etiologically antisera to herpesvirus hominis by indirect he- involved, the data at most suggest an effect magglutination-inhibition. Appl Microbiol of multiple viruses rather than any single 21:680-684, 1971 16. Bernstein MT, Stewart JA: Indirect hemaggluti- entity upon genetically predisposed indi- nation test for detection of antibodies to cyto- viduals. megalovirus. Appl Microbiol 21:84-89, 1971 17. Kaye HS, Ong SB, Dowdle WR: Detection of REFERENCES coronavirus 229E antibody by indirect hemag- 1. Hinden E: Mumps followed by diabetes. Lancet glutination. Appl Microbiol 24:703-707, 1972 1:1381, 1962 18. Steinberg AG, Rushforth NB, Bennett PH: On 2. McCrae WM: Diabetes mellitus following genetics of diabetes mellitus. In Pathogenesis of mumps. Lancet 1:1300-1301, 1963 Diabetes Mellitus. Edited by E Cerasi and R Luft. Stockholm, Almquist and Wiksell, 1970 3. Block MB, Berk JE, Fridhandler LS, et al: "OUTBREAK" OF JUVENILE DIABETES MELLITUS 287 19. Pathogenesis of diabetes mellitus. Br Med J Med 114:500-503, 1963 3:594-595, 1971 26. Craighead, JE : personal communication, 1974 20. Adams SF: The seasonal variation in the onset of 27. Pappenheimer AM, Kunz, LJ, Richardson S: acute diabetes. Arch Intern Med 37:861-864, 1926 Passage of coisackie virus (Conn-5 strain) in 21. Gamble DR, Taylor KW: Seasonal incidence of adult mice with production of pancreatic dis- diabetes mellitus. Br Med J 3:631-633, 1969 ease. J Exp Med 94:45-65, 1951 22. Muntefering H: Zur pathologie des diabetes mel- 28. Coleman TJ, Gamble DR, Taylor KW: Diabetes litus der wei Ben Maus bei der EMC-Virusinfek- in mice after coxsackie B4 virus infection. Br Med tion. Virchows Arch Abt A Pathol Anat J 3:25-27, 1973 356:207-234, 1972 29. Kibrick S, Benirschke K: Severe generalized 23. Boucher DW, Notkins AL: Virus-induced diabe- disease (encephalohepatomyocarditis) occurring tes mellitus. I. Hyperglycemia and hypoin- in the newborn period and due to infection with sulinemia in mice infected with enceph- coxsackie virus, group B. Pediatrics 22:857-875, alomyocarditis virus. J Exp Med 137:1226-1239, 1958 1973 30. Hierholzer JC, Farris WA: Follow-up of children 24. Craighead JE: Workshop on viral infection and infected in a coxsackievirus B-3 and B-4 out- diabetes mellitus in man. J Infect Dis break: no evidence of diabetes mellitus. J Infect 125:568-570, 1972 Dis 129:741-746,1974 25. Craighead JE, Peralta PH, Shelokov A: Demon- 31. Watkins PJ, Soler NG, Fitzgerald MG, et al: stration of encephalomycarditis virus antibody in Diabetic ketoacidosis during the influenza epi- human serums from Panama. Proc Soc Exp Biol demic. Br Med J 4:89-91, 1970

Journal

American Journal of Epidemiology – Pubmed Central

Published: Oct 1, 1974

There are no references for this article.