117 Rheumatic Fever and Rheumatic Heart Disease

Acute rheumatic fever is a noncontagious disease characterized by febrile, nonsuppurative inflamma­tion, primarily of articular and cardiac tissues, less frequently affecting the skin and brain.

Etiology and Treatment

The disease is caused by infection, most often of the throat, with type A beta-hemolytic strains of strepto­coccus. Fever, migratory joint pains and tachycardia, the most frequent symptoms, typically begin 1 to 3 weeks after the onset Ofuntreated streptococcal phar­yngitis. However, only 0.1 to 3.0 percent of untreated bouts of this infection result in a first attack of rheu­matic fever. Consequently, various largely unidenti­fied permissive factors must participate in initiating the immunologic pathogenesis of the disease.

First attacks of acute rheumatic fever can be pre­vented by timely treatment of the streptococcal infec­tion with penicillin or another appropriate antibi­otic, but such treatment does not influence the course of the disease once it has begun. Rheumatic fever recurs only as a result of a new infection with a pathogenic strain of streptococcus. Prophylactic anti­biotic treatment diminishes, but does not eradicate recurrences (Taranta et al. 1964). The shorter the interval since the previous bout of rheumatic fever, the greater is the likelihood that a new attack will be elicited. An infection that occurs within 2 years of an attack has a 20 to 25 percent chance of inducing a recurrence. If the first attack does not affect the heart, a recurrence usually spares it as well, but if the heart has been involved, a second bout it likely to result in greater damage (Spagnuolo, Pasternack, and Taranta 1971). An attack of rheumatic fever usually lasts several weeks, but is rarely fatal.

History

Clinical Manifestations and Diagnosis

The major clinical manifestations of rheumatic fever were first described separately, their relationships not being recognized until the ninteenth century. Thomas Sydenham in 1685 distinguished an acute, febrile polyarthritis “chiefly attacking the young and vigorous” from gout. One year later, he described as “St. Vitus’ dance” the neurological disorder that is now called Sydenham’s chorea (Murphy 1943). It was Richard Bright who in 1839 first connected the condi­tion with rheumatic fever (Schechter 1935).

In 1797, Matthew Baillie of London noted a thick­ening of some heart valves in autopsies of patients who had had acute rheumatism, and a few years later David Dundas, surgeon to King George III, described nine cases of “a peculiar disease of the heart” that “is always the consequence of, or is connected with, rheu­matic affection” (Dundas 1809). Four years later, Wil­liam C. Wells, also of London, published a series of 16 cases of “rheumatism of the heart” (median age, 15 years), and added the description of subcutaneous nodules. However, the occurrence of nodules was largely ignored until a comprehensive study by Thomas Barlow and Francis Wamer was published in 1881 (Benedek 1984).

Before the introduction of auscultation by Ren6 T. Laennec in 1818, rheumatic heart disease was recog­nized from abnormalities of the pulse, respiration, and palpation of the chest in the presence or recent history of fever and joint pains. Laennec had de­scribed murmurs caused by deformities of the mitral valves. In 1835 James Hope also described murmurs that originated from the other valves and concluded that rheumatic fever is the most frequent cause (Hope 1846).

Bacteriologic studies, beginning in the 1870s, were made on blood and joint fluid aspirates. This approach Wasjustified in 1883 by the demonstration of the cause of gonococcal arthritis. The normal bac­terial flora of the throat was not understood. Yet contrary to most investigators of the time, a Berlin physician writing at the turn of the twentieth cen­tury concluded that because one finds the same bac­teria in the throats of persons with or without rheu­matic fever, the disease must be caused not by a specific microbe but by a peculiar reactivity of sus­ceptible individuals.

When Homer F. Swift of New York began his investigations of the cause of rheumatic fever in 1916, he was still culturing blood and joint fluid. In retrospect, the successful cultures that prolonged efforts to find a directly infectious agent (rather than an immunologic incitor) were due either to contamination or to the presence of bacterial endo­carditis, an infection that usually is superimposed on rheumatic or Congential heart disease. By 1928 Swift had concluded that an allergic response to repeated streptococcal infections is the most likely cause of rheumatic fever, but erroneously believed that nonhemolytic strains were the principal patho­gens. The investigations begun by Alvin F. Cobum in New York in the mid-1920s resulted in the cor­rection and extension of Swift’s conclusions. On the basis of detailed epidemiological correlations of the bacterial flora of the throat with the onset of initial and recurrent bouts of rheumatic fever, Coburn in­ferred that the true pathogen is the hemolytic strep­tococcus, serologic type A (Benedek 1987). The hypothesis that the disease is mediated immunologi­cally was supported by the discovery by E.

Early Epidemiological Studies

The first statistical report, by John Haygarth in 1805, found that 4.5 percent of the admissions to the Chester Infirmary between 1767 and 1801 had “rheu­matism” and that 36 percent of those, or 1.6 percent of the total, had acute febrile rheumatism or “rheu- matick fever.” How many of these actually were cases of this disease is uncertain (a problem that continued to vex many surveys). Eleven of the 70 cases of whom age was recorded involved patients 50 to 70 years old, and unequivocal symptoms of heart failure were not described even in fatalities.

Most reports since then have also described the experience in an individual hospital. Regional sur­veys and mandatory reporting of rheumatic fever have been rare, occurring mainly in Great Britain, Scandinavia, and the United States. Nearly a cen­tury after Haygarth’s monograph, a report on the occurrence of “rheumatism” at St. Bartholomew’s Hospital, London, demonstrated the vagaries of diag­nosis in more detail, as well as the great prevalence of rheumatic fever. During the 15-year span 1863 to 1877, an average of 323 patients with “rheumatism” were admitted annually. This constituted about 14 percent of all medical admissions. Beginning in 1867, gonorrheal rheumatism and rhematoid arthri­tis were diagnosed separately. However, this change only slightly reduced the cases of presumed acute rheumatic fever with or without heart disease - to 311 per year. The associated mortality was 1.3 per­cent (Southey 1878). This result is consistent with more recent U.S. data according to which the acute mortality has ranged from 3.5 percent during 1935­42 to 0.9 percent during 1951-8 (Mayer et al. 1963). At St. Bartholomew’s during 1867-77, a total of only 716 cases of scarlet fever (also caused by infection with the hemolytic streptococcus) were admitted, and the mortality among these was 14.0 percent (Southey 1878).

Acute rheumatic fever occurred more frequently in older persons in the nineteenth century than in the more recent preantibiotic period. A question­naire sent to physicians by the British Medical As­sociation during 1882-6 regarding their cases of rheumatic fever indicated the mean age of an at­tack among 647 cases to have been 25.4 years, with 4 percent of the patients over 60 years of age (Whipham 1888). Of 400 cases admitted to Roose­velt Hospital in New York during 1872-82, 7 per­cent of the attacks occurred in patients older than 55. By contrast, 25 percent of first attacks occurred between ages 20 and 25, whereas 49 percent of first attacks and 34 percent of all attacks occurred in the third decade of life (May 1884). The latter is similar to the experience at Bellevue Hospital in New York 40 years later, where, as Table VIIL 117.1 shows, 34.5 percent of the cases of acute rheumatic fever were in the third decade. By the 1930s, however, only 13.8 percent of patients admitted to Philadel­phia hospitals for acute rheumatic fever were in their third decade. Similarly, a review of 3,129 cases in New York, the majority of whom had their first attack between 1920 and 1940, showed a mean age of onset of 15 years (Lambert 1920).

In the 1880s, it was estimated in London that about 20 percent of patients without symptoms or history of rheumatic fever who were admitted to a hospital for any reason had a sibling or parent who had had rheumatic fever; such familial cases were present in about 35 percent of the patients who were hospitalized because of rheumatic fever (Garrod and Table VIII.117.1. Age distribution of admissions for acute rheumatic fever

Sources: Data from “Lambert (1920); and ⁶Hedley (1940).

Cooke 1888). It had been noted in a British Medical Association survey that rheumatic fever “greatly preponderates in the lower over the middle and up­per classes” (Whipham 1888).

The importance of crowded living conditions gained further attention because of the prevalence of the diagnosis of rheumatic fever in military encamp­ments during the First World War. Between April 1917 and December 1919, for example, 24,770 U.S. soldiers, representing 27 percent of all rheumato- Iogic cases, received this diagnosis. Crowded train­ing camps were believed to have contributed to the disproportionately large percentage of cases (Hench and Boland 1946). The development of epidemiolog­ical data about rheumatic fever has been difficult because the acute phase of the disease may be brief or actually imperceptible, so that its occurrence may become recognized only from findings of valvular heart disease, perhaps years later in an uncertain proportion of the cases. Many estimates of the preva­lence of rheumatic fever have been extrapolated from the prevalence of heart disease in those be­tween the ages of 5 and 19 years, because their disease is most likely to be rheumatic. Widely differ­ing diagnostic criteria, examining techniques, and reporting practices have made epidemiological com­parisons unreliable. This problem began to be ad­dressed in 1943 by the Cardiovascular Diseases Sub­committee of the [U.S.] National Research Council, which sought more precise diagnostic criteria so as to assess the problem of rheumatic fever in military training camps. In response, T Duckett Jones, a Boston cardiologist, devised a set of five “major” and seven “minor” criteria and proposed that “any single major manifestation with at least two of the minor manifestations would seem to place the diagnosis on reasonably safe grounds.” The Jones criteria re­ceived international acceptance quite rapidly and have been modified twice. The main revision in 1955 altered the importance that was assigned to some clinical signs, whereas the revision in 1965 placed increased importance on laboratory evidence of re­cent streptococcal infection (Stollerman et al. 1965).

Treatment

Salicylate preparations introduced between 1876 and 1879 were found to be particularly effective in coun­teracting the fever and joint pain of acute rheumatic fever. In 1899 aspirin came on the market, and for the next 50 years the basic treatment consisted of the maximum tolerated dose of either sodium salicylate or aspirin (Rodnan and Benedek 1970). Greater un­derstanding of the cause of rheumatic fever did not have an immediate effect on its treatment because there was no way to eradicate a streptococcal infec­tion. Sulfanilamide was found to prevent recurrences when taken prophylactically (Swift, Moen, and Hirst 1938; Thomas and France 1939). Penicillin became available in 1945 and quickly proved to be safer and more reliable as a preventive agent (Spink et al. 1946). The most reliable prophylactic method, a monthly injection of slowly excreted benzathine peni­cillin, was introduced in 1951.

The introduction in 1949 of cortisone and cortico­tropin was perceived as the first possible improve­ment over salicylate therapy for acute rheumatic fever. During the first few years in which these hor­mones were available, most investigators found that control of fever and joint pain was achieved more rapidly with these drugs than with salicylates; more­over, if given early enough, they seemed to minimize heart damage. Uncertainty about the latter impres­sion, however, led to the establishment in 1955 of an elaborate Anglo-American prospective therapeutic comparison of salicylate with cortisone or cortico­tropin therapy. After 10 years, it was concluded that the long-term results from both modalities were simi­lar (U.K. and U.S. Joint Report 1965).

Decline of Rheumatic Fever

London physician G. B. Longstaff (1905) may have been the first to suggest that the prevalence of rheu­matic fever was decreasing. He deduced this from death certificates in England and Wales during 1881-1900. During these 20 years, 51,666 deaths were attributed to “rheumatic fever” or “rheuma­tism of the heart,” and 3.3 times as many (171,298) to “valvular disease of the heart.” In comparing the 5-year periods 1881-5 and 1896-1900, death rates per million due to rheumatic fever declined 15.6 percent, from 97.6 to 82.4, and those for “residual rheumatism” declined 19.1 percent from 35.0 to 28.3. Data collected by J. A. Glover (1930) for the same region showed a lower incidence of deaths for 1901 than Longstaffhad calculated (67 per million), but indicated continuity of the decline: to 46 per million in 1910 and 38 per million in 1928.

The two most common sources of epidemiological data other than death reports have been hospital admissions and recurrence rates among patients who were being followed in rheumatic fever clinics. In order for hospitalization data to be intrinsically com­parable, one must assume a socioeconomically stable service area and consistent admissions policies, both of which assumptions are risky. Clinic data tend to yield the most detailed information, but if the clinic is effective, its prevalence information cannot be ex­trapolated to communities that lack similar services.

Annual admissions data for rheumatic fever were recorded between 1906 and 1925 for hospitals in Montreal, Boston, New York, New Orleans, and the Panama Canal Zone. A comparison of the proportion of total admissions for rheumatic fever between 1906-8 and 1923-5, as reflected in Table VIΠ.117.2, shows a decline from 59 percent (New York Belle­vue) to 79 percent (New Orleans Charity) (Seegal and Seegal 1927).

Table VIΠ.117.2. Cases Ofrheumatic fever per 1,000 admissions, various hospitals

Source: Data from Seegal and Seegal (1927).

A comparison of the prevalence of rheumatic fever among U.S. military personnel during the two world wars shows that, as serious as the problem was dur­ing World War II, this diagnosis was made in only 10 percent as many army and 38 percent as many navy personnel. The contribution of more precise diagno­sis and prophylactic programs to this improvement is uncertain.

The lengthiest clinical evaluation of rheumatic fever patients was carried out by M. G. Wilson and collaborators (1958) in New York. Beginning in 1916 and continuing in the same area for 40 years, it reported a steady decline in recurrences of rheu­matic fever, unrelated to treatment. During 1921— 43, 25 percent of children 6.to 13 years of age and 6.1 percent of those aged 14 to 20 suffered recurrences, whereas during 1944-56 only 15.1 percent of the younger and 2.8 percent of the older age group were so affected.

Antimicrobial prophylaxis began to be used in 1952. Yet the age-adjusted recurrence rate for the 5 years 1942-6 was 7.9 percent, and for 1952-6 it was 6.1 percent; this difference was not significant. The socioeconomic circumstances of the area served by the clinic had improved during these decades, but no improvement in the recurrence rate was observed in the poorest segment of the population (Wilson, Lim, and Birch 1958).

A study of the economic correlates of the occurrence of primary and recurrent rheumatic fever in Balti­more during 1960-4 also demonstrated a strong cor­relation between low economic status and the oc­currence of the disease, but only among the white subjects. The annual incidence was consistently greater among blacks, with only slight improvement related to higher socioeconomic category, whereas the incidence among whites diminished markedly with higher socioeconomic category. Thus, the an­nual incidence of rheumatic fever among blacks in the lowest fifth on the socioeconomic scale was 56 percent greater than among comparable whites, whereas the difference reached 478 percent when the most affluent fifths were compared (Gordis, Lilien- feld, and Rodriguez 1969).

In weighing the relative importance of rheumatic fever and rheumatic heart disease against other po­tentially lethal problems of childhood, we find that in the United States during 1939-41, the pair still ranked second, behind accidents, as the leading causes of death in the 10- to 14-year age group. There was a significant sex- and race-related gradi­ent, from 11.2 deaths per 100,000 white boys to 17.4 deaths per 100,000 nonwhite girls (Wolff 1951). This amounted to an average of 4,000 deaths per year for that period. But by 1972-4, an average of only 179 deaths were reported annually, and by 1982-4, deaths had declined to 78 cases per year. Similarly, as Table VIII.117.3 indicates, the reported national incidence of rheumatic fever declined from 10,470 cases in 1961 to 2,793 cases in 1971 and 264 cases in 1981.

Geography

Rheumatic fever was thought to be rare in the trop­ics, as judged by observations on English troops and native populations in India in the mid-nineteenth century. Seeming to confirm that view were clinical and autopsy reports, primarily from India and the Malay Peninsula, which started to appear in the 1890s and were still cited in the 1930s (Clarke 1930). But these publications are somewhat perplex­ing, for beginning in 1925, substantial numbers of cases of rheumatic heart disease were being reported from tropical India.

Nevertheless, rheumatic fever has been found to occur less frequently in tropical than in temperate regions despite the fact that poverty, aspects of which appear to facilitate the occurrence of the dis­ease, is in general more prevalent in the tropics. Thus, observations on military personnel that mini­mize the economic factor have shown a lower inci­dence in tropical encampments. For example, the mean incidence of rheumatic fever among white U.S. Army enlisted men stationed in the United States during 1913-25 was 5.00 plus or minus 2.46 per 1,000, whereas the incidence among similar person­nel stationed in the Philippines was 1.74 plus or minus 1.23 per 1,000 (Faulkner and White 1924). Paradoxically, however, during the decade 1914-23 there were slightly more admissions for rheumatic fever to the General Hospital of Manila than to the Johns Hopkins Hospital in Baltimore - 7.90 versus 7.19 per 1,000 medical admissions (Seegal and Seegal 1927).

A north-to-south declining gradient of rheumatic fever within the United States was suggested during the 1920s (Harrison and Levine 1924). This was confirmed by numerous studies, including one in which the prevalence of rheumatic heart disease was compared among Indian children on reservations in Montana, Wyoming, and Arizona (Paul and Dixon 1937). A similar gradient has been reported from China. An altitude gradient has been demonstrated in Kenya for hospitalizations for acute rheumatic fever, and in Mexico for the proportion of rheumatic heart disease among cardiac patients. In both coun­tries there was significantly more disease in the more temperate highlands.

The “classical” manifestations of rheumatic fever were described in northern Europe, and whether its acute manifestations differ in the tropics has long been debated. If one considers, however, the wide variation in the frequency with which various acute manifestations of the disease have occurred in tem­perate regions, as presented in Table VIII.117.3, it seems doubtful that there are consistent biological differences in the presentation of rheumatic fever in the tropics. Some of the reported inconsistencies re­flect selection biases related to cultural attitudes and medical and other resources, whereas others reflect over- or underdiagnosis. For example, the fact that virtually all of the patients in a small series of cases from Nigeria and Uganda had carditis proba­bly indicates that childhood cases of lesser severity were simply not seen rather than nonexistent.

The wide range of the occurrence of “arthritis” in rheumatic fever cases may be attributed in part to the differentiation by some authors between objec­tive signs of joint inflammation and mere arth­ralgia. Nevertheless, there can be striking differ­ences in the occurrence of arthritis, even when rigor­ous diagnostic criteria may be assumed; for example, 42 percent and 71 percent of the patients had arthri­tis in two series of cases at two different times from Toronto. But the extreme variations in the occur­rence of chorea - ranging from 2 percent to 52 percent - are the most inexplicable. Erythema mar­ginatum is rarely described in dark-skinned chil­dren, but this eruption is sufficiently uncommon in fair-skinned individuals that, assuming some under­diagnosis in the former, a significant difference seems unlikely.

The relationship of subcutaneous nodules to car­ditis may indicate real geographic differences in dis­ease expression. Nodules have occurred in 8 to 12 percent of several large series of cases from temper­ate regions (Benedek 1984), whereas they have been less common in others and have consistently been rare in the tropics. In temperate climates nodules are almost always associated with acute carditis and may be predictive of more severe valve damage. Thus, according to the long-term experience of Wil­son and W. Lim (1957) in New York, 3 percent of cases of mitral insufficiency, 9 percent of cases of combined mitral insufficiency and stenosis, and 30 percent of cases of combined mitral and aortic valve damage had exhibited nodules during the acute phase of illness. In tropical areas there is a weaker association between the occurrence of nodules and carditis and a difference in the valvular involve­ment, or at least this seems to be the case of India. There, carditis not only results in permanent valve injury more rapidly and is more likely to be acutely fatal, but also has a peculiar tendency to cause pure mitral stenosis. This has been described repeatedly since 1935. According to the largest reported experi­ence involving 2,050 cases of rheumatic heart dis­ease in New Delhi, 81 percent of whom were below 30 years of age, 43.5 percent had pure mitral stenosis (Wilson et al. 1958). Reports from other countries do not resolve the question of whether this finding represents an ethnic difference or is a result of delayed initial medical care.

Rheumatic Heart Disease and Its Surgical Repair

Wilson and associates evaluated 385 patients up to 49 years of age who had heart murmurs and had had rheumatic fever before the age of 20. Fully 89 per­cent were asymptomatic; 45 percent of the cohort had pure mitral insufficiency; 41.8 percent had com­bined mitral insufficiency and stenosis; 10.3 percent had both mitral lesions and aortic insufficiency; and 2.9 percent had double mitral and double aortic valve defects (Gordis et al. 1969). Of 78 fatal cases from the population of which those just mentioned were among the survivors, 15.4 percent had only mitral insufficiency, 37.2 percent had mitral insuffi­ciency and stenosis, and 47.3 percent also had aortic insufficiency (Magida and Streitfeld 1957).

The surgical treatment of Theumatically damaged heart valves began in the late 1940s when Dwight E. Harken in Boston performed the first successful mitral commissurotomy (Harken et al. 1948). The procedure was improved by Charles P. Bailey of Philadelphia a year later (Bailey et al. 1960). This was “closed” heart surgery. No attempt to repair a damaged aortic valve was possible until an oxygenat­ing system to bypass blood temporarily around the heart was developed; this technique made “open” heart surgery possible. The first practical apparatus was employed by John W. Kirklin. The first valve operations to ameliorate aortic insufficiency were performed by Bailey in 1959 and to correct aortic stenosis in the same year by Donald G. Mulder in Los Angeles. The next technical phase was the re­placement of an active valve. This began with the ball valve devised by Albert Starr and Lowell Ed­wards (1961). Such plastic valves, however, tend to destroy red blood cells, and thus valves were devel­oped with leaflets of pig, cattle, or sheep tissue. This type of prosthesis was first inserted to replace an aortic valve in 1965, and a mitral valve in 1967, both by A. Carpentier in Paris.

In 1983 about 16,000 mitral and 33,000 aortic valve replacements were performed in the United States. However, the cause of the injury, particularly of the aortic valve, has gradually shifted from rheu­matic to other varieties of heart disease (Gillum 1986).

Thomas G. Benedek

Bibliography

Arora, R., et al. 1981. Clinical profile of rheumatic fever and rheumatic heart disease: A study of 2,500 cases. Indian Heart Journal 33: 264—9.

Bailey, C. P., J. Zimmerman, and W. Likoff. 1960. The com­plete relief of mitral stenosis: Ten years of progress toward this goal. Diseases of the Chest 37: 543-60.

Benedek, T. G. 1984. Subcutaneous nodules and the differentiation of rheumatoid arthritis from rheu­matic fever. Seminars in Arthritis and Rheumatism 13: 306-21.

1987. A century of American rheumatology. Annals of Internal Medicine 106: 307—12.

Clarke, J. T. 1930. The geographical distribution of rheu­matic fever. Journal of Tropical Medicine and Hygiene 33: 249-57.

Dimdas, D. 1809. An account of a peculiar disease of the heart. Medico-Chirurgical Transactions 1: 36—46.

Faulkner, J. M., annd P. D. White. 1924. The incidence of rheumatic fever, chorea and rheumatic heart disease. Journal of the American Medical Association 83: 425-6.

Garrod, A. E., and E. H. Cooke. 1888. Frequency of rheu­matic family histories amongst non-rheumatic pa­tients. Lancet 2: 110.

Gillum, R. F. 1986. Trends in acute rheumatic fever and chronic rheumatic heart disease: A national perspec­tive. American Heart Journal 111: 430—2.

Glover, J. A. 1930. I. The incidence of acute rheumatism. Milroy lectures on the incidence of rheumatic fever, 1: 499-505.

Gordis₁ L., A. Lilienfeld₁ and R. Rodriguez. 1969. Studies in the epidemiology and preventability of rheumatic fever. Journal of Chronic Diseases 21: 645—66.

Harken, D. E., et al. 1948. The surgical treatment of mitral stenosis. I. Valvuloplasty. New England Jour­nal OfMedicine 239: 801—9.

Harris, T. R., and S. A. Levine. 1924. Notes on the regional distribution of rheumatic fever and rheumatic heart disease in the United States. Southern Medical Jour­nal 12: 914-15.

Hedley, O. F. 1940. Rheumatic heart disease in Philadel­phia hospitals, part II. Public Health Reports 55: 1647-91.

Hench, P. S., and E. W. Boland. 1946. The management of chronic arthritis and other rheumatic disorders among soldiers of the United States Army. Annals of Internal Medicine 24: 808-25.

Hope, J. 1846. A treatise on the diseases of the heart and great vessels, and on the affections which may be mis­taken for them, 2d American edition, from 3d London edition, 95-113. Philadelphia.

Lambert, A. 1920. The incidence of acute rheumatic fever at Bellevue Hospital. Journal OftheAmerican Medical Association 74: 993-5.

Longstaff₁ G. B. 1905. A contribution to the etiology of rheumatic fever. Transactions of the Epidemiological Society of London, new ser., 24: 33-83.

Magida, M. G., and F. H. Streitfeld. 1957. The natural history of rheumatic heart disease in the third, fourth and fifth decades of life. II. Prognosis with special reference to morbidity. Circulation 16: 713-22.

May, C. H. 1884. Statistics of four hundred cases of rheu­matism, with special reference to treatment: treated at the Roosevelt Hospital. Medical Record 25: 57-62, 87-93, 116-21, 173-8.

Mayer, F. E., et al. 1963. Declining severity of first attack of rheumatic fever. American Journal of the Diseases of Childhood 105: 146-52.

Murphy, G. E. 1943. The evolution of our knowledge of rheumatic fever. Bulletin of the History of Medicine 14: 123-47.

Paul, J. R., and G. L. Dixon. 1937. Climate and rheumatic heart disease: A survey among American Indian school children in northern and southern localities. Journal of the American Medical Association 108: 2096-100.

Rodnan, G. P., and T. G. Benedek. 1970. The early history of antirheumatic drugs. Arthritis and Rheumatism 13: 145-65.

Schechter, D. C. 1975. St. Vitus’ dance and rheumatic disease. New York State Journal of Medicine 75: 1091-102.

Seegal, D., and B. C. Seegal. 1927. Studies in the epidemiology of rheumatic fever. Journal of the American Medical Association 89: 11-17.

Southey, R. 1878. Observations on acute rheumatism. St. Bartholomew’s Hospital Reports 14: 1—22.

Spagnuolo, M., B. Pastemack, and A. Taranta. 1971. Risk of rheumatic fever recurrences after streptococcal in­fections. NewEngland Journal OfMedicine 285:641-7.

Spink, W. W., et al. 1946. Sulfadiazine and penicillin for hemolytic streptococcus infections of the upper respira­tory tract. Archives OfInternal Medicine 77: 260—94.

Starr, A., and M. L. Edwards. 1961. Mitral replacement: Clinical experience with a ball-valve prosthesis. An­nals OfSurgery 154: 726-40.

Stollerman, G. H., et al. 1965. Jones criteria (revised) for guidance in the diagnosis of rheumatic fever. Circula­tion 32: 664—8.

Swift, H. F., J. K. Moen, and G. K. Hirst. 1938. The action of sulfanilamide in rheumatic fever. Journal of the American Medical Association 110: 426—34.

Taranta, A., et al. 1964. Rheumatic fever in children and adolescents. IV. Relationship of the rheumatic fever recurrence rate per streptococcal infection to the titers of streptococcal antibodies. Annals of Internal Medicine 60 (Supplement 5): 47-57.

Thomas, C. B., and R. France. 1939. A preliminary report of the prophylactic use of sulfanilamide in patients susceptible to rheumatic fever. Bulletin of the Johns Hopkins Hospital 64: 67-77.

Todd, E. W., 1932. Antigenic streptococcal haemolysin. Journal OfExperimental Medicine 55: 267—80.

U.K. and U.S. Joint Report. 1965. The natural history of rheumatic fever and rheumatic heart disease: Co­operative clinical trial of ACTH, cortisone, and aspi­rin. Circulation 32: 457-76.

Vendsborg, P., L. F. Hansen, and K. H. Olesen. 1968. De­creasing incidence of a history of acute rheumatic fever in chronic heart disease. Cardiologia 53: 332—40.

Whipham, T. 1888. Reports of the Collective Investiga­tion Committee of the British Medical Association. III. Acute rheumatism. British Medical Journal i: 387-404.

Wilson, M. G., and M. Schweitzer. 1954. Pattern of heredi­tary susceptibility in rheumatic fever. Circulation 10: 699-704.

Wilson, M. G., and W N. Lim. 1957. The natural history of rheumatic heart disease in the third, fourth and fifth decades of life. I. Prognosis with special reference to survivorship. Circulation 16: 700—12.

Wilson, M. G., W N. Lim, and A. M. Birch. 1958. The decline of rheumatic fever: Recurrence rates of rheu­matic fever among 782 children for twenty-one con­secutive calendar years (1936-1956). Journal of Chronic Diseases 7: 183-97.

Wolff, G. 1951. Death toll from rheumatic fever in child­hood. Journal of the American Medical Association 145: 719-24.