Recurrent acute rheumatic fever attacks, disabling in themselves, pose an increased risk for development of or exacerbation of rheumatic heart disease. For those who have had rheumatic fever, the American Heart Association [4] recommends a single injection of benzathine penicillin G every 4 weeks or daily, oral administration of penicillin V, sulfadiazine, or erythromycin to prevent intercurrent streptococcal infections. Although few patients receive lifelong antistreptococcal prophylaxis, no general consensus exists as to when prophylaxis can be safely discontinued.

The risk for acute rheumatic fever recurrence, however, is neither continuous nor uniform. It declines with the age of the patient and the number of years since the most recent attack. It is positively correlated with the number of previous attacks and with the presence and severity of preexisting rheumatic heart disease [5, 6]. In recognition of these facts and because of limited public health resources, the Chilean National Health Service has established guidelines for prophylaxis of acute rheumatic fever that focus on the persons at highest risk. Duration of prophylaxis is also constrained by a heavy reliance on benzathine penicillin G, due to the poor compliance with oral prophylaxis in the patients treated by the National Health Service. This policy results in cessation of prophylaxis in a number of adolescents and young adults judged on clinical and epidemiologic grounds to be at lower risk for acute rheumatic fever recurrence. We did prospective studies of streptococcal infection and acute rheumatic fever recurrences among a group of such patients residing in the Southeast Health District of Santiago, Chile.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

All study patients were enrolled in the Rheumatic Fever Control and Prevention Program of the Southeast Health District, conducted by the Department of Public Health at Catholic University Medical School, Santiago, Chile. The Health District's approximately 750 000 inhabitants are predominantly low income and low-middle income, and 80% are cared for by the national health system.

Patients were entered in the Control and Prevention Program because of an attack of acute rheumatic fever (observed and documented by the Program physicians) that fulfilled the modified Jones criteria or because they were referred from ambulatory clinics with rheumatic valvular heart disease. The latter was confirmed by Program cardiologists who did clinical examinations as well as reviewed electrocardiograms and chest roentgenograms.

Criteria for Cessation of Prophylaxis

On admission to the Program, all nonpenicillin-allergic patients were given prophylaxis with injections of benzathine penicillin G every 28 days. Those with a history of penicillin allergy received oral sulfadiazine, 1 g daily. Patients who had not had carditis during their acute attack were maintained on continuous antimicrobial agent prophylaxis for 5 years or until age 18, whichever was longer. Patients with carditis during the acute attack who had no detectable cardiac sequelae were kept on prophylaxis for 10 years or until age 25, whichever was longer. The same regimen was used for patients whose sole cardiac sequela was mild mitral regurgitation. Those with aortic valvular involvement, mitral stenosis, or polyvalvular disease received prophylaxis for life.

Study Protocol

All patients in the present study had completed their prescribed period of continuous antimicrobial agent prophylaxis (n = 48), had declined to continue intramuscular penicillin (n = 9), or were allergic to penicillin (n = 2). Patients in the latter two categories were enrolled only after it was apparent that they were noncompliant with oral sulfadiazine prophylaxis. They were, however, continued under prospective surveillance, consisting of regular clinic visits every 3 months when they were interviewed and examined by a study nurse. In addition, a throat culture and blood sample were obtained. Patients who had any signs or symptoms suggestive of rheumatic fever were referred to a cardiologist for further evaluation. Patients were encouraged to report promptly to the study clinic if they developed symptoms suggestive of acute pharyngitis or rheumatic fever. Cultures were obtained from those with sore throats and, if cultures were positive for group A streptococci, intramuscular benzathine penicillin G (or oral erythromycin in penicillin-allergic individuals) was administered. No treatment was given for asymptomatic group A streptococcal throat carriage detected by routine study surveillance cultures. In addition to the regular quarterly visits, each patient had an annual examination by the study cardiologist, including an electrocardiogram and a chest roentgenogram.

Laboratory Tests

Throat swabs were incubated overnight in 5 mL of Todd-Hewitt broth, then a loopful of broth was plated on 5% sheep blood agar. Plates were incubated for 18 to 24 hours at 37 °C in room air. Colonies of ß-hemolytic streptococci were subcultured to obtain pure cultures and were grouped either by capillary precipitation using BBL (Cockeysville, Maryland) antisera or by agglutination with Streptex reagent (Burroughs-Wellcome Co., Research Triangle Park, North Carolina). All ß-hemolytic streptococci were sent to the laboratory of one of us (ALB) where serogrouping was reconfirmed. All group A isolates were analyzed for M and T serotypes by standard methods [7, 8] using antisera provided by the Centers for Disease Control. Antistreptolysin O titers were assayed at the Catholic University Public Health Unit Streptococcal Laboratory using commercially available reagents (BBL), as were serum titers of anti-DNAse B (Streptonase B, Wampole Laboratories, Cranbury, New Jersey). All sera from a given patient were tested in a single run at the conclusion of the study. Dilution increments were approximately 0.15 log, and increases of two tubes (0.3 log), representing a twofold increase in serum titer, were considered clinically significant.

Statistical Analysis

Ninety-five percent confidence intervals were determined using the Poisson probability function generator (SAS version 6.03, SAS Institute Inc., 1988).

Results

Top Methods Results Discussion Author & Article Info References

Patient Follow-up

The study began in July 1982 and terminated in September 1988. During this period, 59 postprophylactic patients (approximately 80% of whom entered the study during the first 6 months) were followed prospectively for a total of 1032 scheduled visits and 3346 patient-months (mean months of follow-up per patient, 56.7 months; median, 67.5 months; range, 10 to 75 months). The patients ranged in age at entry from 15 to 44 years; they comprised 40 women and 19 men. Fifteen (38%) of the women, but only two (10%) of the men had rheumatic heart disease (Table 1). Ten patients (17%) had a history of multiple attacks. Only six patients were 5 years or less from their most recent rheumatic attack (Table 2).

There were 34 unscheduled visits, of which 25 were due to the complaint of sore throat. Throat cultures were positive for ß-hemolytic streptococci in 10 instances, but only 2 were positive for group A streptococci. In one patient treatment with benzathine penicillin G was administered, and no recurrence of acute rheumatic fever occurred. The case history of the second patient (Patient 046) is described below.

Twelve patients entered in the off-prophylaxis protocol did not complete the study. In two instances this was due to recurrence of acute rheumatic fever (see below). Two patients originally thought to be allergic to penicillin were subsequently found able to tolerate the drug and were placed back on prophylaxis. One patient left Santiago, and seven patients were lost to follow-up. The mean period of follow-up of these patients was 29.2 months (range, 9 to 52 months).

Culture and Serologic Test Results

Throat cultures and streptococcal antibody assays were done routinely only during the period from July 1982 through September 1986. During this period, 756 patient visits and 2393 patient-months of follow-up occurred (mean months of follow-up per patient, 40.5 months; median, 46.2 months; range, 5 to 51 months).

Serogroups of 124 ß-hemolytic streptococci isolated (including scheduled and unscheduled visits) were as follows: A, 29 strains; B, 35; C, 24; F, 1; G, 14; nongroupable, 21. Thus the isolation rate of group A streptococci was 1 per 82.5 patient-months of follow-up (equivalent to 14.5 [CI, 10.1 to 20.8] per 100 patient-years). The 29 group A strains were isolated from 19 patients, and the strains had a wide variety of T-agglutination patterns. Only two were M typable: one strain each of M4 and M6. Group A streptococci were isolated from 9 of 33 (27%) patients, ages 15 to 24; 10 of 18 (56%) patients, ages 25 to 34; and none of 8 patients, 35 or older. Nine of the 29 (31%) group A streptococcal isolations were associated with an increase in antibody titer.

Increases in antibody titer were detected in 56 instances: antistreptolysin O alone, 26; anti-DNAse B alone, 16; antistreptolysin O plus anti-DNAse B, 14. Because of lapses in follow-up, serologic data indicating immunologically significant streptococcal infection were based, in a number of instances, on sera collected more than 3 months apart. The intervals between samples showing increases were 3 months or less, 39; 4 months, 11; 5 months, 4; and 6 months, 2. The rate of serologically documented streptococcal infection was 28.1 (CI, 21.7 to 36.5) per 100 patient-years. Thirty percent of the increases in both antistreptolysin O and anti-DNAse B titers were four tubes in magnitude or greater (Figure 1).

View larger version (29K): [in this window] [in a new window]   Figure 1. Antibody increases. Frequency and magnitude of increases in antibody titers and their relation to isolation of groups A, C, and G streptococci. Note that the horizontal axis (number of tubes increase) refers to the magnitude of the increase in tube dilution increments. ASO = antistreptolysin O. ADNASE = anti-DNAse B.

Eight of the 56 increases in antibody titer were associated with positive throat cultures for group A streptococci at the initial or follow-up visits or both. In five of these eight instances, increases of both antistreptolysin O and anti-DNAse B titers occurred.

Six increases were associated with isolations of group C streptococci and three increases were associated with group G organisms. Seven of these nine increases were of antistreptolysin O alone (all two tubes in magnitude), and two were of anti-DNAse B alone (two tubes and three tubes) (see Figure 1).

Seasonal distribution of the increases in antibody titer was assessed by the date of the serum showing the increase and by omitting pairs collected more than 4 months apart. Results were summer, 16; fall, 12; winter, 14; spring, 8.

A history of recent sore throat or fever or both was elicited in association with 13 of the 56 serologic increases, 15 of the 29 instances of recovery of group A streptococci from the pharynx, 5 of the 8 increases associated with positive cultures for group A streptococci, and 4 of the 9 increases accompanied by isolations of group C or G streptococci.

Recurrences of Rheumatic Fever

There were two recurrences of rheumatic fever during the 3346 patient-months (278.8 patient-years) of follow-up between 1982 and 1986, yielding a recurrence rate of 0.72 (CI, 0.2 to 2.6) per 100 patient-years of prospective surveillance. Details of the recurrences are described below.

Patient 048

A woman had attacks of Sydenham chorea at ages 5, 7, and 13 years. During her second attack, a mitral insufficiency murmur was heard, but mitral valve prolapse was subsequently diagnosed from the echocardiogram. At age 19 years, in April 1984, she had another recurrence of chorea without any other major clinical manifestations of rheumatic fever. Group A streptococci were not isolated from her pharynx. Her antistreptolysin O titer was 1:333, but this did not represent a significant increase from recent serial determinations. Prophylaxis was reinstituted after the recurrence.

Patient 046

A woman had acute rheumatic fever with polyarthritis, subcutaneous nodules, and mild carditis in March 1978 at age 24 years. She was lost to follow-up until her fifth pregnancy in 1982. The patient had no clinical evidence of rheumatic heart disease when she successfully completed this pregnancy nor did she on her annual evaluation in 1983. She steadfastly refused benzathine penicillin G injections but received continuous follow-up examinations. In July 1983, she returned to the clinic on an unscheduled visit with fever, sore throat, and a positive throat culture for group A streptococci (T, 11; M, nontypable). She was treated with erythromycin, 1.5 g per day, for 10 days. One month later on a scheduled visit, her throat culture was again positive for the same serotype of group A streptococcus, and the treatment was repeated. During this period her antistreptolysin O titer increased from 1:50 to 1:166, and the anti-DNAse B titer increased from 1:340 to 1:680. Throat cultures in December 1983 and subsequent cultures were negative for group A streptococci. Mitral stenosis was diagnosed on a routine clinic visit in June 1984 at age 30 and was subsequently confirmed by echocardiography. At this time, she agreed to benzathine penicillin G prophylaxis and has had no recurrences of acute rheumatic fever. Although the development of mitral stenosis might have been a consequence of her original rheumatic attack 6 years before, we have classified her as a recurrence for our analysis.

Epidemiology of Acute Rheumatic Fever in the Community

Acute rheumatic fever continued to occur in the Southeast Health District of Santiago during the study. Between 1982 and 1985, an average of 22.5 patients per year were diagnosed. The incidence increased markedly in 1986, when 52 case patients were diagnosed. In that year, the attack rate of acute rheumatic fever among children ages 6 to 14 years in the Health District, reached 21.7 per 100 000 population. Thirty patients were diagnosed during 1987 and 26 were diagnosed during 1988. Neither of the two recurrences in study patients occurred during this time. Since the termination of the study, a gradual decline has occurred in the number of diagnosed patients with acute rheumatic fever in the Health District, with only six patients being diagnosed in 1991.

Discussion

Top Methods Results Discussion Author & Article Info References

Previous Studies

Three previous prospective studies Table 3 have analyzed the risk for recurrence in patients with a history of acute rheumatic fever who were not receiving continuous antistreptococcal prophylaxis. The protocols used in those three studies were generally similar to our protocol, except that follow-up was at 2 rather than 3 months.

Johnson and colleagues [9, 10] followed 233 adolescents and 125 adults in Chicago for 906 patient-years. All of the patients were at least 5 years from their most recent attack, and more than one half had heart disease. The acute rheumatic fever recurrence rate was 1.3 per 100 patient-years for adolescents and 0.7 for adults.

Feinstein and colleagues [11] studied Irvington House patients who had reached the ages of 14 to 25 years and who were at least 3 to 5 years from their most recent rheumatic attack. Patients with clinical evidence of rheumatic heart disease were excluded. Patients received either a placebo or 200 000 units of potassium penicillin G daily by mouth. After a total of 278 patient-years of follow-up, the attack rates per 100 patient-years were 1.4% (two recurrences) in the placebo group and 0.7% (one recurrence) in the penicillin group, a difference that was not statistically significant (P > 0.5). This study differed from the other investigations in that symptomatic streptococcal throat infections were not treated. Thus the observed recurrences may have been preventable.

Bisno and colleagues [12] in Memphis, Tennessee, maintained a group of older adults off prophylaxis and withdrew a second group thought to be at low risk for recurrence (age 16 years or older, at least 5 years since the most recent attack, and free of rheumatic heart disease) during the summer months. No recurrences occurred during 106 patient-years of surveillance.

Recurrence Rates

The rate of recurrence in our patients from Santiago (0.7 per 100 patient-years) was low and remarkably similar to those observed in prospective studies done in the United States at Irvington House and in Chicago some 2 decades earlier, in which recurrence rates ranged from 0.7 to 1.4 (see Table 3). Of the two recurrences in Santiago, one was in a patient with multiple attacks of Sydenham chorea. We previously reported [13] that it is extremely difficult to prevent such recurrences in patients from Santiago even with diligent benzathine penicillin G prophylaxis.

It is difficult to compare the rates of streptococcal infection in all these studies (10-12, this study) because of methodologic differences in the definition of infection, the number and types of antibody tests done, and whether symptomatic infections were treated with antibiotic agents. Moreover, increases in antistreptolysin O titer not associated with isolations of group A streptococci could be due to infections by group C or group G strains, both of which elaborate antistreptolysin O. Strains of these latter two serogroups are commonly isolated from the throats of adults, have caused severe pharyngitis in common-source outbreaks associated with contaminated food or water [14], and are suspected of causing sporadic cases of sore throat [15, 16]. Indeed, some group G strains have M proteins that are similar functionally and structurally to those of Streptococcus pyogenes [17, 18].

Despite the methodologic difficulties mentioned above, the recurrence rates of acute rheumatic fever per 100 streptococcal infections have been rather similar (3.5 to 6.7) in all studies in which adolescent and adult patients were removed from prophylaxis and in which recurrences occurred (see Table 3). This is considerably lower than the rate of 16.9 in children aged 5 to 15 years in the original Irvington House study [1]. Thus, the lower recurrence rate in adolescents and adults may not only be due to a decreased exposure to streptococcal infections but also to a decreased propensity to reactivate rheumatic fever once infection has occurred. This evidence for a decreased propensity was shown by Spagnuolo and colleagues [6], who did a multivariate analysis and found that young age was an independent risk factor for recurrence of acute rheumatic fever after streptococcal infection.

The incidence of acute rheumatic fever is known to vary greatly both by geographic location and time period. This phenomenon might be related in part to the characteristics of the prevalent streptococcal strains. Strains epidemiologically associated with epidemics of rheumatic fever tend to belong to a relatively limited number of M serotypes [19, 20] that have common structural features [21], fail to synthesize the -lipoproteinase known as opacity factor [22], and are often heavily encapsulated [23, 24]. Thus the incidence of acute rheumatic fever in the community, reflecting partly the presence of rheumatogenic strains, might increase the risk for recurrence. A recent example of this phenomenon has been reported from West Virginia [25]. It is noteworthy that the low recurrence rate of acute rheumatic fever observed in our carefully selected patients occurred in the presence of continuing rheumatic activity in the community and even during an outbreak of the disease. Little evidence suggests that our patients were exposed to highly rheumatogenic strains. The percentage of strains that was M typable was remarkably low, only 7% as compared with 52% of throat isolates in our previous studies in this group of patients [26]. Only one strain belonged to a recognized rheumatogenic serotype (M6), and none of the isolates grew as mucoid colonies.

Discontinuing Prophylaxis

What can be concluded from our own study and those conducted in the 1950s and 1960s in the United States? Although the data are admittedly limited, they indicate that individuals who have reached their early twenties, had their most recent attack more than 5 years ago, did not have carditis with their previous attack(s), and are free of rheumatic heart disease can be taken off prophylaxis with relative safety. An acute rheumatic fever recurrence rate of approximately 1 attack/100 patient-years can be anticipated without prophylaxis, and, given the mimetic nature of rheumatic recurrences [27, 28], most recurrences will not involve the heart.

Some authorities [28] would be even more conservative and would treat patients with Sydenham chorea similarly to those with healed carditis even if no evidence of heart involvement exists. The rationale for this approach is that, given the long latent period between the antecedent streptococcal infection and the onset of chorea [29, 30], one cannot exclude the possibility that the patient had occult carditis that had resolved before the onset of neurologic symptoms. Moreover, the insidious development of rheumatic heart disease has been documented in some patients years after an episode of chorea [31, 32]. Because episodes of chorea occur in prepubescent children, in most instances continuation of prophylaxis until the early twenties would approximate the current Chilean criteria for patients with healed carditis.

Despite the previous guidelines, a decision to discontinue rheumatic fever prophylaxis must be individualized and should be taken only after consultation with the patient regarding the relative risks and benefits. Prudence dictates that certain factors should be taken into account. These factors include the likely exposure of the patient to children or adolescents with streptococcal infections in the home or work place, his or her reliability in reporting promptly to the physician any episodes of pharyngitis, and the incidence of rheumatic fever in the community.

Patients with rheumatic heart disease are at greater risk for recurrence [5, 6, 10] and are more likely to have carditis should a recurrence ensue. According to Chilean National Health Service policy, patients with healed carditis or with rheumatic heart disease who had only mild mitral regurgitation received prophylaxis until age 25 or 10 years after their most recent rheumatic attack, whichever was longer. In our study, no recurrences occurred among such patients after completing the prescribed period of prophylaxis. The one presumed recurrence was in patient 046, who refused prophylaxis.

The published data ([10, 12], this study) on discontinuing acute rheumatic fever prophylaxis in patients with rheumatic heart disease, however, are too limited to allow definitive recommendations at this time. We shall continue to monitor our patients carefully and adhere to the guidelines outlined above. Additional studies of the issue from diverse geographic locations are clearly warranted.