The efficacy of drug desensitization is best understood in IgE-mediated immediate hypersensitivity. In addition, it is also effective for nonallergic immediate reactions, such as infusion-related reactions, which directly trigger mast cells to secrete their granules or cytokines [27].

Two categories of desensitization protocols are currently available: RDD and slow drug desensitization (SDD). RDD is recommended for immediate reactions, both allergic and nonallergic. The most widely used RDD protocol is doubling the dosage every 15 minutes until the therapeutic dose is achieved. SDD is recommended for type IV delayed hypersensitivity reactions with T cell involvement, and can be performed both orally and intravenously. There is as yet no consensus on SDD protocols, including the initial dose, dose increments between steps, and dosing interval. Further clinical experience and research are required to establish the role and efficacy of desensitization for delayed reactions [28,29].

Before performing drug desensitization, the diagnosis should be confirmed by thorough history taking and testing. Desensitization should be considered when no alternative options are available or the culprit drug has far superior efficacy, i.e., the benefits of the culprit drug outweigh the risk. It should be carefully applied in patients with significant risk factors for fatal BTRs, such as those with a history of severe hypersensitivity reactions, like anaphylaxis, Stevens-Johnson syndrome, and toxic epidermal necrolysis, those with cardiovascular or respiratory diseases, and those using beta-blockers. The appropriate protocol should be selected based on the concentration, dosage, dosing interval, and route of administration (oral, intravenous, intramuscular, or subcutaneous). Oral and intravenous protocols are widely used because intramuscular and subcutaneous protocols involve frequent injections. The route of administration can be switched without compromising the efficacy. It is feasible to perform desensitization through intravenous drug administration for meticulous dose adjustment and administer the therapeutic dose in oral form.

H1 blockers, H2 blockers, and glucocorticoids can be used as premedication. Aspirin and montelukast block the end products of the arachidonic acid cascade and decrease the incidence and severity of BTRs. NSAIDs can help to control the symptoms of cytokine release syndrome. Glucocorticoids alone are not recommended because they cannot prevent the initial degranulation of mast cells [30].

BTRs should be promptly managed depending on their severity. For mild to moderate BTR, desensitization should be stopped, and appropriate medication, such as antihistamines and/or glucocorticoids, should be administered to the patient. After the symptoms have resolved, desensitization can be resumed at the same step or the step prior to that in which BTR occurred, and most patients succeed in completing the protocol without further BTRs. In cases of anaphylaxis, intramuscular epinephrine, massive hydration, and proper oxygenation are essential. Desensitization should be stopped until the patient is stabilized, and further desensitization protocols should be tailored by an allergy specialist.

Immunological tolerance can be maintained with continued administration of the drug. For drugs that are administered intermittently, such as chemotherapeutic agents, desensitization should be performed at every cycle, as the tolerance fades during the drug-free interval. If the patient undergoes multiple cycles of desensitization, the number of steps can be gradually tapered unless BTRs occur. However, shortening of the protocol should be performed carefully by a specialist, considering the severity of the initial DHR, the characteristics of the drug itself, as well as the frequency, pattern, and degree of BTR.

Antineoplastic agents are repeatedly administered at regular time intervals, which increase the risk of sensitization to these drugs. The incidence of DHR to antineoplastic agents varies according to the type of drug, ranging from 1% to 80% [31–33]. The common causative agents include platinum (e.g., cisplatin, carboplatin, and oxaliplatin), taxanes (e.g., paclitaxel and docetaxel), and monoclonal antibodies (e.g., rituximab, infliximab, trastuzumab, bevacizumab, etc.) [34]. DHR is commonly caused by sensitization with drug-specific IgE, which can be confirmed by skin tests [35–38]. However, non-IgE-mediated reactions, such as infusion-related or solvent-related reactions (e.g., cremophor, polysorbate 80), have also been reported in the literature [32,33,39]. Mast cell-mediated or basophil-mediated reactions present with urticaria, angioedema, laryngeal edema, vomiting, abdominal pain, or anaphylaxis. On the other hand, cytokine-mediated reactions present with generalized rash, abrupt chest discomfort, dyspnea, back pain, fever, chills, or myalgia.

As mentioned above, the most prevalent causes of DHR include platinum, taxanes, and monoclonal antibodies, which present with different clinical manifestations. DHR to platinum has been reported in 1% to 44% of patients [30]. By gradual sensitization after multiple exposures, the first DHR most frequently occurs on the 8th to 9th administration, especially when it is resumed after a drug-free period. The correlation between skin test positivity and DHR has been well established. As cross-reactivity can occur between platinum agents, care should be taken when switching agents within platinum-based regimens. As BTRs are common even after previous successful desensitization, shortening of the protocol should be adopted gradually and carefully. In contrast, infusion-related reactions to taxanes tend to occur during the first or second administration. In addition, IgE-mediated reactions can occur in some patients who have specific IgE to the pollen of Taxus baccata, a common yew used as a source of taxanes. DHR to taxanes has been reported in 5% to 10% of patients. As taxane-associated BTRs usually occur during the initial attempts of desensitization, the number of steps within desensitization protocols can be safely reduced if the BTR does not occur repeatedly during the early cycles of desensitization. Monoclonal antibodies lead to various types of DHR, including IgE-mediated reactions, cytokine release syndrome, and mixed reactions, which can arise in both the earlier and later courses. The rate of DHR varies according to the degree of humanization and glycosylation pattern, and commonly reported agents include rituximab and cetuximab. Especially for cetuximab, the association with anti-α-1,3-galactose IgE has been observed, which is generated from the consumption of animals that were bitten by a specific type of tick [30].

Various desensitization protocols have been proposed for antineoplastic agents. A protocol from the Adverse Drug Reaction and Desensitization Program in Brigham and Women’s Hospital has been well established worldwide (Table 1) [36]. In this protocol, the culprit drug solution (1:1) is diluted to 1:10, 1:100, and sometimes 1:1,000, and the concentration and rate of administration is escalated stepwise for 15 minutes. This multi-bag protocol is completed in most patients without serious BTR [40–42]. However, multi-bag protocols require excessive time and effort for dilution and changing bags, and increase the risk of exposure of medical staff to antineoplastic agents in the process. In addition, the stability of the drug can be compromised by dilution and administration for a prolonged period.

To overcome these shortcomings, a one-bag protocol was proposed in which a single concentration was used and only the rate of administration was increased (Table 2) [43–46]. Successful completion of desensitization with a one-bag protocol has been reported for all platinum, taxane, and monoclonal antibodies. There were no statistically significant differences regarding the rate of completion or the incidence of BTRs compared to multi-bag methods, while the amount of time required for the preparation was significantly reduced. In addition, the incidence of BTR decreased in the later course of treatment, which was consistent with multi-bag protocols [43–46]. Therefore, it is anticipated that the one-bag protocol can replace the multiple-bag protocol, thus reducing time and effort, maintaining the stability of the agents, preventing erroneous dilutions, and reducing the risk of exposure of medical staff to antineoplastic agents.

In addition, the efficacy of chemotherapy is unlikely to be compromised by desensitization. Sloane et al. [42] conducted a large retrospective analysis of patients with recurrent ovarian cancer who underwent RDD for carboplatin; overall survival in the RDD group was non-inferior to that of the regular administration group.

DHR to β-lactams, such as penicillin or cephalosporin, is more common than that to non-β-lactams. Desensitization can be performed for both immediate and delayed hypersensitivity reactions. The protocol should be selected based on patient characteristics, hospital capacity, and physician preferences. It is generally started with 1/1,000 of the therapeutic dose and then increased by 2 to 3-fold every 15 minutes to 5 hours. Oral administration is preferred due to its ease, safety, and effectiveness [47]. Desensitization to penicillin and cephalosporins has been well established. Successful desensitization has also been reported for other β-lactams, such as carbapenem and monobactam, and non-β-lactams, such as vancomycin, clindamycin, metronidazole, macrolides, aminoglycosides, tetracycline, and ciprofloxacin [48–58].

Red man syndrome (RMS) is a distinctive hypersensitivity reaction to vancomycin, ciprofloxacin, teicoplanin, or amphotericin; it is caused by the direct activation of mast cells and secretion of histamine, and is associated with the rate of administration [59]. It can usually be prevented by reducing the rate of administration and pretreating patients with antihistamines. However, in cases of moderate-to-severe RMS, the desensitization protocol can be applied in the absence of appropriate alternative agents. Both RDD and SDD have been studied for RMS; in most cases, RDD can be applied first and, in the case of failure, the protocol can be switched to SDD [60,61].

Successful desensitization to other antimicrobials has also been reported for antifungals, such as amphotericin B, fluconazole, itraconazole, voriconazole, and micafungin, and for antivirals, such as acyclovir, valganciclovir, ribavirin, and nevirapine [62–69].

Tuberculosis possesses unique microbial characteristics that require prolonged simultaneous administration of multiple agents. DHR to antituberculous agents is reported in 4% to 5% of patients, which leads to the discontinuation or alteration of medications or prolonged treatment duration [70]. It is common for patients with tuberculosis to be hypersensitive to one or two agents, and delayed reactions, such as maculopapular rash, are more prevalent than immediate reactions, such as urticaria or angioedema [71].

Successful cases of desensitization to antituberculous agents, both oral and intravenous, have been reported in the literature since 1990. Through desensitization, 60% to 80% of patients who were hypersensitive to first-line antituberculous agents could continue therapy with those drugs (Table 3) [72]. In addition, second-line agents, such as streptomycin, kanamycin, and cycloserine, have also been continued safely with intramuscular, intravenous, and oral desensitization protocols [73]. In a Korean multicenter study of patients with severe cutaneous adverse reactions to antituberculous agents, the replacement of all drugs did not prevent the recurrence of DHR, possibly due to the development of multi-drug hypersensitivity after drug reaction with eosinophilia and systemic symptoms [74]. However, 84.6% of patients undergoing desensitization succeeded in continuing the original drugs, and their BTRs were no greater than mild.

Both RDD and SDD were studied at intervals of 15 minutes to 1 hour for RDD and a few hours to a day for SDD [75]. RDD has strengths in patient compliance and medication resistance, while SDD has better success rates and safety [29,72,76]. SDD has been recommended for delayed DHR; however, multiple successful RDDs for delayed reactions have also been reported [29].

Hypersensitivity reactions to aspirin/NSAIDs can have varying manifestations, depending on the onset of symptoms and the organs involved [77]. Various protocols for aspirin/NSAIDs, which are adapted from aspirin provocation test protocols, have been suggested [3,78–82]. The mechanism of aspirin/NSAID hypersensitivity is still not fully understood, but non-immunological mechanisms of cyclooxygenase (COX) inhibition are the most plausible. The desensitization process downregulates the production of cysteinyl leukotriene and its receptors, attenuating skin and airway responses [83].

Aspirin/NSAID desensitization is considered for patients with cardiovascular or musculoskeletal diseases who require aspirin or NSAID administration for prolonged periods. It is also considered for patients with aspirin/NERD to improve the overall symptoms and quality of life, and decrease the formation of nasal polyps and sinus infections; however, the effects on asthma outcomes are less consistent [84].

Oral desensitization protocols for aspirin are well established. Notably, the therapeutic dose at the final step differs according to the indication for aspirin/NSAIDs. For cardiovascular or musculoskeletal diseases, a lower dose of aspirin (75 to 100 mg/day) is used. In contrast, for aspirin-exacerbated respiratory disease or chronic sinusitis with or without nasal polyps, a higher dose of aspirin (up to 625 mg/day) is administered (Table 4) [84].

For DHR to NSAIDs other than aspirin, it is recommended to switch to drugs that have little COX-1 inhibitory effect (e.g., acetaminophen), drugs that exhibit relatively selective COX-2 inhibition (e.g., nimesulide and meloxicam), or drugs that are highly selective COX-2 inhibitors (e.g., celecoxib and etoricoxib) if well tolerated. The desensitization protocol for NSAIDs can be modified from that for aspirin [77]. The temporary tolerance to aspirin/NSAIDs lasts 48 to 72 hours after desensitization. Therefore, hypersensitivity reactions can recur 2 to 5 days after discontinuation if the therapeutic dose is not continued [85,86].