This editorial refers to ‘Turning 18 with congenital heart disease: prediction of infective endocarditis based on a large population’†, by C.L. Verheugt et al., on page 1926
Infective endocarditis (IE) has remained a serious problem in cardiology as neither the incidence nor the mortality of this disease has decreased over the last decades.1 Despite major diagnostic and therapeutic advances, prognosis has remained poor and mortality high.2 The epidemiological profile of IE has changed over the years, particularly in industrialized countries.1 Whereas younger patients with previously known, mostly rheumatic valve disease were primarily affected in the past, the number of older patients who more often develop IE as the result of healthcare-associated procedures with either no previously known valve disease or with degenerative disease with or without valve replacement has now dramatically increased.3,4 Within the group of younger patients the epidemiological profile has also changed. While affected patients frequently had rheumatic valve disease in the past, congenital heart disease (CHD) is now predominating.2 The population of adults with CHD, on the other hand, has markedly increased over the past decades and is currently estimated at >1.2 million in Europe.5 Despite major advances in the medical care of these patients, a variety of possible complications remain frequent during long-term follow-up.6,7 IE is one of the most serious events in this population (Figure 1).8 The incidence of IE ranges from 3 to 10 episodes/100 000 person-years in the general population.2,3 The proportion of CHD in patients with IE has been reported to be between 2 and 18%,9–11 with a consistent minor male predominance.8,12,13 This variation in numbers is probably due to selection bias. Importantly, the reported incidence of IE in those with CHD is 15–140 times higher than in the general population.2,12,14 Again selection bias presumably causes this wide variation. The incidence seems to vary for the different defects, and the highest estimate originates from a highly specialized centre with a higher selection of complex cases. Some simple lesions, such as secundum atrial septal defect and pulmonary valve stenosis, appear to carry a particularly low risk of IE.2 However, reports of IE incidence for the various congenital defects have in general to be viewed with caution as lesions frequently occur in combination. Ventricular septal defect is, for example, often considered as a high risk defect; a prospective study of these patients, however, suggests that IE is primarily related to associated valve disease rather than to the septal defect itself.15 In summary, it has to be emphasized that our knowledge of IE in patients with CHD is still limited as systematic studies are few, often retrospective, and heavily affected by selection bias.
Although prognosis of IE is better in congenital than in acquired heart disease, mortality remains high, with a reported range between 4 and 10%.8,12,13 Thus, prevention of this serious complication is one of the major tasks in the care of adults with CHD. Although surgical repair of CHD may reduce the risk of IE provided there is no residual lesion and no non-endothelialized artificial material,10,16 many patients may require prosthetic material for repair and minor residual defects are relatively common. In this situation, the overall risk of IE may still be markedly increased after repair.2 Thus, there is no scientific justification for a repair with the sole purpose of eliminating the risk of IE.17 Besides general preventive measures, antibiotic prophylaxis has been widely practised for many years. The approach to antibiotic endocarditis prophylaxis has, however, recently changed for several reasons.2,18 In short, transient bacteraemia occurs not only after dental procedures but frequently in the context of daily routine activities such as tooth brushing, flossing, or chewing. Due to the lack of scientific evidence for the efficacy of antibiotic prophylaxis, the estimated huge number of patients that may need to be treated to prevent one single case of IE, the small but existing risk of anaphylaxis, and the general problem of emergence of resistant microorganisms resulting from widespread, often inappropriate use of antibiotics, it is currently recommended by expert consensus to limit antibiotic prophylaxis to patients with the highest risk of IE undergoing the highest risk procedures.2,18 This recommendation includes the following patient groups: patients with a prosthetic valve or a prosthetic material used for cardiac valve repair; patients with previous IE; and certain patients with CHD, i.e. (i) cyanotic CHD, without surgical repair, or with residual defects, palliative shunts, or conduits; (ii) CHD after repair with prosthetic material whether placed by surgery or by a percutaneous technique, up to 6 months after the procedure (until endothelialization); and (iii) when a residual defect persists at the site of implantation of a prosthetic material or device following cardiac surgery or a percutaneous technique.2,18 The recommendation has furthermore been limited to dental procedures requiring manipulation of the gingival or periapical region of the teeth or perforation of the oral mucosa. Antibiotics are no longer recommended for respiratory tract, gastrointestinal, genitourinary, dermatological, or musculoskeletal procedures unless there is an established infection.2,18 The updated recommendations have dramatically changed long-established practice for primary care physicians, cardiologists, dentists, and their patients. They have therefore raised a great deal of discussion particularly among physicians who take care of patients with CHD.
Against this background, Verheugt et al.19 provide important new information. The group must be congratulated for establishing such a large national registry of adults with CHD. This database allowed the study of—among many other aspects—the occurrence of IE in a population of presently 10 210 patients. With an incidence rate of IE of 1.1 per 1000 patient-years, the study confirmed the increased risk in these patients compared with 1.7–6.2 per 100 000 patient-years in the general population.3 The study also confirmed the particularly low risk of patients with pulmonary stenosis and atrial septal defect as long as these conditions are not associated with additional lesions. The highest risk was found in pulmonary atresia with ventricular septal defect [hazard ratio (HR) 16.7] followed by other complex lesions such as congenitally corrected transposition of the great arteries and univentricular heart (HR 7.1 and 6.1), but also simple lesions such as bicuspid aortic valve and ventricular septal defect (HR 6.3 and 6.8). Predictive complications during childhood were IE and cerebrovascular events. Based on the available data, a prediction model was derived and transformed into a clinically applicable risk score. Up to age 40, patients with a low predicted risk (<3%) had an observed incidence of <1%; those with a high predicted risk (≥3%) had an observed incidence of 6%. The model also yielded accurate predictions up to age 60.
Is such a prediction model likely to provide individually tailored medical surveillance and educational counselling in the near future as suggested by the authors? Not only would the model need to be confirmed in other patient cohorts, but there are also important limitations of the study that need to be considered before use in clinical practice. The registry does not exclude selection bias. Patients have been recruited and included through the treating cardiologist or via response to advertisements in the local media. More importantly, critical data are missing in the database: surprisingly, there was no significant difference between corrected and uncorrected defects. This could have been due to too small patient numbers in subgroups (there was a difference that did not reach statistical significance). Maybe more importantly, residua may cause increased risk after repair, and data on residua were unfortunately not available for analysis. Data on traditional risk factors such as preceding dental procedures, intravenous drug abuse, tattoos, piercing, or the use of antibiotic prophylaxis could have added important information but were unfortunately also not available.
Thus, for the time being, the results of this study may not change the approach to IE in CHD.
Patients with CHD and their caring physicians must always be aware of the risk of IE. Any clinical suspicion must prompt careful evaluation including echocardiography and blood culture in order to provide early diagnosis and to allow effective treatment.
As far as preventive measures are concerned, it has to be emphasized that good oral hygiene and regular dental review have an essential role in reducing the risk of IE. Aseptic measures are mandatory during manipulation of venous catheters and during any invasive procedure in order to reduce the rate of healthcare-associated IE. CHD patients should also be discouraged from getting piercings and tattoos.18
As far as antibiotic prophylaxis is concerned, the currently recommended restriction of its use to the patients with the highest risk undergoing the procedures with the highest risk is supported rather than conflicted by the presented data. Whether the use should be extended to patient groups and procedures not included in the current recommendations (see above2,18) will require further research. Careful prospective documentation of risk factors for IE, use of preventive measures, and occurrence of IE in large unselected CHD patient populations must be encouraged to improve our knowledge of IE and appropriate care further in the setting of CHD.
Conflict of interest: none declared.