Penicillin Reactions in Patients With Severe Rheumatic Heart Disease: A Presidential Advisory From the American Heart Association

Serious adverse events following BPG administration are frequently attributed to (presumably penicillin) anaphylaxis because of the association with a known allergen and rapid progression to collapse. Although some BPG reactions are the direct result of anaphylaxis, publications dating back >50 years make it clear that not all severe BPG reactions in patients with RHD are anaphylactic.⁹ Recently, a growing number of anecdotal reports¹⁰ of immediate or nearly immediate BPG‐related deaths, many of which did not have features of classical anaphylaxis, prompted a case review.⁶ In that series, 10 cases from 5 countries were reported. Only 3 met criteria for possible anaphylaxis; the other 7 cases showed remarkable similarity: A patient with severe valvular RHD who lost consciousness almost immediately after BPG injection and could not be resuscitated.⁶ Similarly, a review of the Khartoum RHD registry recently reported on >800 children with RHD.¹¹ In that series, 3 young children, over a span of only a few months, experienced collapse and immediate death following BPG injection, despite receiving appropriate intervention for presumed anaphylaxis. No anaphylaxis symptoms, other than collapse, were noted, but all 3 had severe mitral regurgitation with resultant heart failure.

These data suggest that cardiovascular compromise, rather than anaphylaxis, may be responsible for some sudden deaths that occur immediately after BPG injection, in particular in patients with severe valvular heart disease. Marantelli and colleagues,⁶ as part of the case review previously mentioned, outlined the physiological cascade of responses that may precipitate these events. Although the mechanisms vary slightly based on the distribution of underlying structural heart disease, it is postulated that pain or fear of BPG administration drives a physiological response precipitating decompensation that includes vasovagal hypotension, bradycardia (in some cases), and syncope, quickly leading to decreased coronary perfusion, ventricular arrhythmias, and sudden cardiac death.

Although there is no direct evidence yet to support this mechanism of death, we believe it is highly plausible and supported by 4 arguments. First, there is a consistent association between adverse outcomes after BPG injections and severity of RHD.^{6, 11, 12} Anaphylaxis, compared with cardiac compromise, would show equal distribution across severity of RHD proportionate to the number of BPG injections delivered, yet sudden deaths occur disproportionately in young people with severe, symptomatic RHD. Second, there is a clear discrepancy of adverse BPG events between patients receiving BPG for RHD prophylaxis and those receiving BPG for other indications. Most data are from women receiving BPG to prevent mother‐to‐child transmission of syphilis, with few reports of adverse BPG reactions, despite a large number of doses of BPG being given.^{13, 14, 15, 16} Third, there has been an apparent lack of response to appropriate anaphylaxis treatment in many patients, which also casts doubt that anaphylaxis is the cause of death. Although the patient response to resuscitation efforts, including administration of adrenaline, is not part of any diagnostic criteria for anaphylaxis, in clinical practice, adrenaline is highly effective in reversing immediate symptoms. However, the 2019 case series of adverse drug reactions to BPG found that although 8 of 9 patients received adrenaline, 7 were unable to be resuscitated. Finally, there has consistently been a striking lack of other anaphylactic symptoms among patients with severe RHD who experienced sudden death. Although hypotension and loss of consciousness can be features of an anaphylactic death, they are most often preceded by tachycardia along with respiratory, cutaneous, or gastrointestinal manifestations, which have been notably absent from these reports.

The following are clinical features comparing vasovagal collapse with anaphylaxis¹⁷:

The most important factor predisposing patients to cardiac compromise appears to be the severity of the underlying valvular heart disease. Patients with severe valve disease, regardless of the valve in question, have little cardiovascular reserve and may not compensate well to pain on injection, vasovagal syncope, or volume and pressure shifts associated with administration of BPG. Adrenaline administration, although appropriate and lifesaving for anaphylaxis, may contribute to further cardiovascular compromise precipitated through tachycardia and decreased time for ventricular filling. Patients with severe mitral stenosis who depend on increased preload to maintain cardiac output may be at the highest risk, followed by those with severe aortic stenosis and severe aortic insufficiency. Additionally, many of the reported deaths have occurred in patients with severe mitral regurgitation in the setting of symptomatic heart failure and left ventricular dysfunction.¹⁸ Thus, we believe that while well‐compensated severe mitral regurgitation does not have increased risk and is not a contraindication to BPG, symptomatic patients with severe mitral regurgitation, in particular those who are cachectic and have decreased left ventricular systolic function, are at elevated risk. Because patients with severe MR often straddle (or move back and forth between) low‐ and elevated‐risk categories, this group requires frequent individualized assessment of risk, with adjustment regarding oral versus injectable penicillin for patients as needed.

In many settings, we believe that there are also modifiable environmental conditions that may contribute to increased risk in vulnerable patients. Vagal reactions are likely exacerbated by conditions more common in low‐ and middle‐income countries where the majority of fatal BPG reactions have been reported. Often, patients presenting to injection clinics for BPG in low‐ and middle‐income countries appear dehydrated or hungry after having traveled long distances, waiting hours for injections, and sometimes having experienced a recent intercurrent illness—explaining why patients who have tolerated injections in prior months might have a reaction without significant change in their underlying disease. In addition, many receive injections in settings where routine hemodynamic monitoring is unavailable, making it more challenging to distinguish anaphylaxis from a cardiovascular collapse.

Based on the available evidence presented in this advisory, the expert working group believes patients with RHD who have severe valvular heart disease or heart failure may be at elevated risk of cardiovascular compromise and sudden death following BPG administration. In addition, the benefit of secondary prophylaxis is questionable given that the purpose of secondary prophylaxis is to prevent progression to severe valve lesions. Thus, we believe the risk of adverse reactions to BPG in patients at elevated risk outweighs its theoretical benefit. For these patients at elevated risk, we newly advise that oral prophylaxis, if reliably available and affordable, should be strongly considered (Table 1). We recognize in making this recommendation that there are availability concerns associated with oral prophylaxis in some settings, which will need to be considered. In many cases, regular prescribing of oral long‐term prophylaxis would require a commitment from government in order to ensure availability. Additionally, although unequivocally inferior, reducing the risk of acute rheumatic fever recurrence by 87% compared with 96% with BPG, high adherence to oral penicillin provides reasonable protection against recurrent group A beta‐haemolytic streptococci infection (71% versus 91% with BPG),¹⁹ and eliminates the risk of injection‐related vasovagal syncope, which may lead to increased mortality in this elevated‐risk group. Guidelines for oral prophylaxis can be drawn from American Heart Association (AHA)⁵ or World Health Organization (WHO)²⁰ recommendations or from local guidelines and best practices, where they exist. Adherence support for those taking oral prophylaxis deserves special attention (at least twice‐daily dosing) and should be tailored as needed to promote high adherence.

If BPG is preferred, we advise that, if possible, it be given to patients at elevated risk in a monitored setting by individuals experienced in cardiopulmonary resuscitation and proper cardiopulmonary resuscitation equipment. As no standardized severity scale for RHD exists, utilization of the American College of Cardiology/AHA classification for valvular heart disease selecting for the valve most severely affected is a reasonable approach for classifying patients at elevated risk.²¹

For patients with less than severe valvular heart disease and preserved ventricular function, we believe there is a yet to be understood but low risk of cardiovascular compromise should a vasovagal reaction occur. However, even given this presumed low risk, we advocate for a multifaceted strategy for vasovagal risk reduction in all patients with RHD receiving BPG, as most of these recommendations are simple, can be implemented in most locations, and have the potential to save lives. However, at the same time, we emphasize that the inability to institute all of these recommendations in an individual setting should not preclude BPG administration. As current guidelines recommend, all low‐risk patients without a history of penicillin allergy or anaphylaxis should be prescribed BPG for secondary prophylaxis given its superior efficacy in the prevention of recurrent rheumatic fever.¹⁹

For all patients receiving BPG, we emphasize that standard best practice should include reducing injection pain and patient anxiety, both of which are known risk factors for injection‐related syncope.²² A range of pain reduction strategies exists for BPG injections including firm pressure to the site for 10 seconds before injection, application of an ice pack before injection, and the use of simple analgesics such as paracetamol before the injection.²³ In addition, patients should be well‐hydrated before injection. Consumption of 500 mL of water has been shown to prevent reflex syncope by increasing vascular constriction and blood pressure, with cardiovascular effects peaking ≈30 minutes after ingestion and lasting for 60 minutes, although clincians should modify this recommendation as appropriate when considering the patient with reduced ventricular function.^{24, 25, 26, 27, 28} We also advise that patients eat at least a small amount of food within 60 minutes before receiving an injection.^{29, 30} Finally, acknowledging that patient and clinician preferences may impact this decision, we advise administration of BPG in patients in the supine position, which limits venous pooling in the lower extremities, and remaining supine for 5 to 10 minutes following injection with slow return to the standing position (Table 2).

We also believe that health care professionals who administer BPG should be taught to recognize vasovagal symptoms (sweating, lightheadedness, nausea, ringing in the ears, blurred or reduced vision, sudden feeling of hot/cold, pallor, dilated pupils, yawning) and should educate their patients as appropriate. Patients with presyncopal symptoms should lie back down immediately and leg elevation should be used, if possible, to increase cerebral vascular flow. A cool damp cloth or ice pack applied to the face and neck may further attenuate symptoms and avoid loss of consciousness. Although symptoms can often be reversed through early initiation of counterpressure maneuvers (isometric contractions such as hand grip or leg crossing), which can increase cerebral blood flow velocity and cerebral oxygenation,^{31, 32, 33} they are generally contraindicated in patients with heart failure and in those with critical valvular stenosis and, thus, should be reserved for patients in the low‐risk category.

Although the assessment of patients for penicillin allergy is largely outside the scope of this advisory, we wish to acknowledge that BPG skin testing, using dilute BPG before injection, has been widely practiced in many parts of the world including African countries, Iran, and Nepal.³⁴ It is possible that the practice stems from the 2001 WHO guidelines on acute rheumatic fever and RHD, which state that: “health workers need to be trained on skin testing before giving BPG injection for secondary prophylaxis.”³⁵ Despite this widespread practice, there is no evidence that skin testing with dilute BPG is useful, and it cannot be recommended.

However, formal skin testing of patients who have a history of hypersensitivity to penicillin or allergy to penicillin has a high negative predictive value for anaphylaxis. Formal testing involves subcutaneous injection of benzylpenicilloyl polylysine (major determinant) penicillin G diluted with normal saline to 10 000 units/mL (minor determinant), positive and negative controls.³⁶ This test is not expected to be readily available in primary or secondary health care settings in limited‐resource countries, and, in these settings, clinical history most commonly identifies those labeled as being allergic to penicillin. However, even when formal skin testing is available, it is not recommended as routine for most patients with RHD who have no history of penicillin sensitivity.

This advisory has been written based on an identified urgent and potentially modifiable clinical situation, BPG‐triggered cardiac deaths, with the hopes of saving lives and bolstering confidence in BPG administration for low‐risk patients around the globe. However, there is little contemporary evidence on adverse BPG reactions, anaphylactic or otherwise, in patients with RHD. Although there are many studies that could be done on this topic, the priority is research to understand the extent and nature of the problem, and to identify and implement solutions. Table 3 contains a summary of outstanding issues considered to be of high importance with identified research priorities.

This expert advisory panel continues to advocate that BPG is the most effective form of secondary prophylaxis and should be prescribed to all low‐risk patients with RHD who have no contraindication to penicillin. However, we acknowledge with this advisory that there is now a growing body of evidence that patients with RHD who have severe valvular heart disease with or without reduced ventricular function may be dying from cardiovascular compromise following BPG injections. Given this emerging data and the questionable benefit of secondary prophylaxis in severe RHD, we strongly advise the prescription of oral antibiotic prophylaxis, preferably oral penicillin, for patients with RHD at elevated risk, if readily available. Further, when possible, we recommend vasovagal risk reduction for all patients with RHD receiving BPG, in the hopes of saving more lives and avoiding syncope events that, even if reversible, can have devastating effects on patient and clinician confidence in BPG (Table 4).