Hypersensitivity and desensitization to antineoplastic agents: outcomes of 189 procedures with a new short protocol and novel diagnostic tools assessment


  • Edited by: Werner Aberer


Ricardo Madrigal-Burgaleta, Hospital Universitario Ramon y Cajal, Servicio de Alergia, Ctra Colmenar Viejo km 9, 100. 28034 Madrid, Spain.

Tel.: +34 913368625

Fax: +34 913368624

E-mail: rmadbur@hotmail.com



Desensitization to antineoplastic agents is becoming a standard of care. Efforts to establish and improve these techniques are being made at many institutions. Our aims are to evaluate a new rapid desensitization protocol designed to be shorter (approximately 4 h) and safer (reducing hazardous drugs exposure risks) and to assess the oxaliplatin-specific immunoglobulin E (IgE) as a novel diagnostic tool.


Prospective, observational, longitudinal study with patients who, for a 1-year period, suffered reactions to antineoplastic agents and were referred to the Desensitization Program at Ramon y Cajal University Hospital (RCUH). Patients were included or excluded as desensitization candidates after anamnesis, skin testing, risk assessment, and graded challenge. Specific IgE was determined in oxaliplatin-reactive patients. Candidate patients were desensitized using the new RCUH rapid desensitization protocol.


Of 189 intravenous rapid desensitizations, 188 were successfully accomplished in the 23 patients who met inclusion criteria for desensitization (of 58 referred patients). No breakthrough reactions occurred in 94% of desensitizations, and most breakthrough reactions were mild. In 10 oxaliplatin-reactive patients, 38 desensitizations were successfully accomplished. Sensitivity for oxaliplatin-specific IgE was 38% (0.35UI/l cutoff point) and 54% (0.10UI/l cutoff point); specificity was 100% for both cutoff points.


In the hands of a Desensitization Program, managed by drug desensitization experts, this new protocol has proven an effective therapeutic tool for hypersensitivity to several antineoplastic agents (oxaliplatin, carboplatin, paclitaxel, docetaxel, cyclophosphamide, and rituximab); moreover, it improves safety handling of hazardous drugs. We report the first large series of oxaliplatin desensitizations. Oxaliplatin-specific IgE determination could be helpful.

Hypersensitivity reactions (HSR) to antineoplastic agents are an increasing important problem. Severity ranges from mild symptoms to life-threatening anaphylaxis, and even death [1, 2]. For example, reactions to carboplatin and oxaliplatin have been reported with incidences from 12 to 17%, and more than 50% of the reactive patients develop moderate to severe symptoms [1, 3, 4]. Many patients presenting with HSR are doomed to therapy discontinuation for fear of inducing more severe reactions [1], even before the disease becomes refractory to treatment; so, survival prognosis is jeopardized [5]. For these reasons, HSR are recognized as important contributors to disease, as impediments to the best treatment of various conditions, including neoplastic maladies [6].

The need to offer first-line therapy to the increasing number of patients who have suffered a HSR has spurred the improvement of rapid desensitizations [7]. This procedure allows patients to receive medications to which they have previously suffered HSR with few or no side effects while escalating to the full therapeutic dose in hours [7]. In the last years, rapid desensitization has become a standard of care in some institutions for safely administering first-line therapies to drug-allergic patients [8].

The primary outcome of this study is to evaluate a new rapid desensitization protocol (designed to be shorter, applicable to several drugs, personalized, and to have definite safety advantages) as a therapeutic tool after protocolized diagnostic approach (including anamnesis, skin testing, risk assessment, and graded challenge).

The secondary outcome is to assess oxaliplatin-specific immunoglobulin E (IgE) as a novel diagnostic tool.

The additional outcome is to assess our practical approach to antineoplastics HSR (a specific Desensitization Program focused on safest access to best treatments for our patients).


Informed consent statement

This study was approved by the Ramon y Cajal University Hospital (RCUH) Ethics Committee (institutional register number: 273/12), which validated informed consents that needed signing by patient, allergist, and referring oncologist.

Patient population

We carried out a prospective, observational, longitudinal study with the patients who, between May 2010 and May 2011, had suffered HSR to antineoplastic agents in our Hospital and had been referred to the Allergy Division Desensitization Program.

Initial reaction

We collected signs and symptoms during initial reactions as recent articles [1]. We included ‘fever/chills’, for it has been reported by other authors [9, 10]. Initial reactions were classified into two groups according to their elapsed time from drug administration to HSR (≤48 or >48 h). Moreover, they were classified in five severity groups as proposed by the National Cancer Institute (NCI) Common Toxicity Criteria [11].

Risk assessment

Patients were divided into two groups (high or low risk). We took on account risk assessment in previous publications [1, 12]. We considered high-risk patients those who met any of the following criteria: patients with previous severe reaction (such as a history of intubation and cardiovascular collapse), patients with comorbidities where exposure might provoke situations beyond medical control (such as uncontrolled asthma or lung disease with forced expiratory volume <1 l in 1 s, or unavoidable use of β-blockers), and pregnant women.

Allergological study: skin testing

Skin tests were performed (by trained personnel in adequate settings) after an interval which allowed for the resolution of symptoms and clearance from the circulation of anti-allergic medications [13], at least 2 weeks after initial reaction to minimize false-negative results [1]. Skin testing was undergone according to previous studies [9, 13-15]. Nonirritant concentrations used by our group are shown in Table 1. Doses for monoclonal antibodies were empirically obtained: full-strength solution according to manufacturer instructions (maximal concentration) and diluted further in normal saline to 1 : 10 (minimal concentration) [9]. We avoided the use of doxorubicin and liposomal doxorubicin because of cutaneous toxicity [16]. Whenever skin tests were positive, endpoint titration was undergone to determine whether the desensitization protocol needed adjusting to a different starting dose [17]. Whenever skin tests were positive and the patient underwent desensitization, skin tests were repeated immediately after the desensitization procedure.

Table 1. Nonirritant concentrations used for skin testing with antineoplastic agents in patients referred to our Desensitization Program [9, 14, 15]
DrugsPrick test concentration (mg/ml)Intradermal test concentrations (mg/ml)
  1. Drugs were prepared by the Pharmacy Department Cytotoxic Unit. For skin prick test, a drop of the drug at the maximal concentration was applied to the volar surface of the forearm by the use of a skin prick test lancet. If prick test was negative, for intradermal injections, 0.03 ml of the drug (first at minimal concentration; and, if negative, maximal concentration) was injected. A positive reaction was defined as a wheal with a diameter at least 3 mm larger than that produced by a negative control (dilutor). Histamine (10 mg/ml) was used as a positive control [1].

Carboplatin101 and 10
Cisplatin10.1 and 1
Oxaliplatin50.5 and 5
Paclitaxel61 and 6
Docetaxel101 and 10
Rituximab101 and 10
Infliximab101 and 10
Cetuximab50.5 and 5
Cyclophosphamide101 and 10
Ifosfamide101 and 10

Allergological study: graded challenge

Patients in the low-risk group with negative results in their skin tests were subjected to graded challenge. It was undergone administering the drug according to the manufacturer instructions, β-blockers were held prior to the procedure, and it was undergone in Medical Intensive Care Unit (MICU) beds assigned to the program [12]. Graded challenge was considered positive when it reproduced the original symptoms [12] or showed an objective HSR.

Allergological study: Gold Standard

Patients with positive skin tests or positive-graded challenge were classified as patients with a positive allergological study.

Patients with a negative-graded challenge were classified as patients with negative allergological study.

Allergological study: oxaliplatin-specific IgE

Serum examples were collected from all oxaliplatin-reactive patients. These examples were sent to ThermoFisher Scientific Phadia AB in Uppsala, Sweden, for oxaliplatin-specific IgE determination by ImmunoCAP®. We calculated sensitivity and specificity for oxaliplatin-specific IgE (cutoff points in 0.10 and 0.35 UI/l) using our allergological study, Gold Standard, for the assessment.

Candidates for rapid desensitization

We considered candidate patients for rapid desensitization those who met these inclusion criteria: (i) patients who had suffered symptoms compatible with immediate type reactions during the drug infusion (or within 48 h); (ii) who had indication, by their referred oncologist, to be treated with the culprit drug; (iii) who signed the informed consents; and (iv) who showed either a high-risk categorization or a positive allergological study for the culprit drug.

Patients with late reactions occurring beyond 48 h after infusion, severe immunocytotoxic reactions, vasculitis, bullous skin diseases like Stevens–Johnson syndrome (SJS)/toxic epidermal necrolisis (TEN), or drug reaction with eosinophilia and systemic symptoms (DRESS)/drug-induced hypersensitivity syndrome (DIHS) were excluded [1, 17].

For candidate patients, all following treatments with the culprit antineoplastic were administered by means of rapid desensitization [6, 7, 18].

Administration of the RCUH rapid desensitization protocol

Hazardous drugs handling guidelines propose a variety of ways to reduce unnecessary occupational exposure to these agents, such as priming intravenous lines with a compatible nondrug solution before spiking the antineoplastic agent bag; and so, hospitals may implement this specific measure to prevent nursing staff from priming lines with hazardous agents in inadequate environments [18]. However, an average line containing around 20 ml of a nondrug solution interferes with the possibility of administering the small volumes planned for the Brigham and Women's Hospital (BWH) standard chemotherapy rapid desensitization protocol [1, 6, 7, 9, 14, 17, 19-24]. The alternative option of priming lines with the antineoplastic itself was considered unacceptable due to the risk of spills and personnel/patient contamination. We modified the previously published protocols developed by the BWH to overcome this problem and designed a new rapid desensitization protocol (Table 2).

Table 2. Ramon y Cajal University Hospital rapid desensitization protocol for a total dose of 200 mg of oxaliplatin (diluted in 250 ml of 5% glucose) originally meant to be infused in 2 h at a 125 ml/h rate
Total dose200 mgSolution concentrationTotal dose in each solution (mg)Drug
Solution A250 ml0.016 mg/ml4.0Oxaliplatin
Solution B250 ml0.160 mg/ml40.0Oxaliplatin
Solution C250 ml0.677 mg/ml169.2Oxaliplatin
StepSolutionRate (ml/h)Administered volume (ml)Time (min)Administered dose (mg)Cumulative dose infused (mg)
  1. Total infusion time: 255 min (4,25 h).

  2. Not all the volume in solutions A and B is infused. Solution C is administered completely. Total dose in solution C is calculated by subtracting the cumulative dose administered in steps 1–8 from the total desired dose.

  3. Whenever needed because of a positive skin test, additional solutions with lower concentrations than solution A may be added previous to solution A to ensure a starting dose as determined on the basis of an endpoint titration.

  4. Solutions were prepared by the Pharmacy Department Cytotoxic Unit and then spiked at bedside by the specialized nursing staff with an individual infusion line previously primed with 22 ml of the dilutor substance. Steps 1, 5, and 9 are considered ‘flushing steps’ (in which 22 ml of dilutor substance are infused).

  5. Step 10 may be adapted to the desired final infusion rate according to the standard regimes indicated by the referred oncologist (additional steps may be added to reach higher infusion rates while maintaining a dose increasing by 2-fold to 2.5-fold with each step).


Infusion pumps with automatic multistep infusion options (Alaris® SE double channel) were used to avoid human errors associated with manually changing infusion rates every 15 min. We used 22 ml infusion systems for these pumps (Alaris® SE I pump smart site infusion set).

Only standard premedications for each drug (according to prescribing information by the manufacturer and institutional protocols) were used. No additional premedication was used for the procedure. Whenever possible, ‘β-blockers were held for 24 h prior to desensitization, because of their possible interference with the therapeutic effect of epinephrine’ [1].

All desensitizations were conducted in MICU beds assigned to the Desensitization Program.

Breakthrough reactions during desensitizations were divided into five groups according to NCI Common Toxicity Criteria [11]. Management of these breakthrough reactions is explained in Table 3.

Table 3. Management of breakthrough reactions during desensitization following local guidelines, adapted from international guidelines (28, 29) and classified as proposed by the National Cancer Institute Common Toxicity Criteria [11]
Grade 1 Mild reactionGrade 2 Moderate reactionGrade 3 Severe reactionGrade 4 Life-threatening; disabling reaction
  1. Antihistamines: dexchlorpheniramine 5 mg along with ranitidine 50 mg.

  2. Corticosteroids: hydrocortisone 200 mg.

  3. All this medications were readily available at bedside. Once the breakthrough reaction was controlled, the protocol was reinitiated from the step it had been paused.

  4. Grade 1 (mild) reactions are mainly cutaneous (flushing or rash, drug fever <38°C), transient, and self-limited (they do not need medication, and stopping the infusion is enough for them to be controlled).

  5. Grade 2 (moderate) reactions include those reactions that do not present with severe symptoms like hemodynamic changes, but do need drug administration to be stopped (rash, flushing, urticaria, dyspnea, drug fever >38°C, moderate rhinitis).

  6. Grade 3 (severe) and grade 4 (life threatening; disabling) reactions stand for reactions that need urgent treatment, and either stand for anaphylaxis or may quickly evolve to anaphylaxis (grade 3: symptomatic bronchospasm, with or without urticaria; parenteral medication(s) indicated; allergy-related edema/angioedema; hypotension. Grade 4: anaphylaxis).

  7. Grade 5 stands for death due to the reaction, and it is not considered in this table.

Immediately stop the drug infusion, check vital constants, call allergist and MICU personnel.
No medication necessary. 30 min observation time required to guarantee patient does not develop a more severe reaction.Consider administer intravenous therapy with antihistamines along with corticosteroids.In case of anaphylaxis symptoms, administer first and promptly: 0.5 ml of adrenaline (1 mg/ml) intramuscularly.

Consider administer intravenous therapy with antihistamines along with corticosteroids.

Consider administer, whenever necessary: oxygen and nebulized bronchodilators.

If a patient who had suffered a breakthrough reaction during desensitization required a new desensitization, the protocol was administered unaltered, but adding pretreatment with acetylsalicylic acid 500 mg orally (instead of 325 mg, because of commercial availability) and montelukast 10 mg orally administered 2 days before the desensitization, based on previous experience [8]. Other measures like decelerating dose escalation with intermediate infusion steps or administering prophylactic medication prior to reactive steps were considered if necessary [1].

Effectiveness and safety assessment

To assess the effectiveness of this protocol, we took on account the number of procedures in which the expected final dose of the drug was administered. We also took on account the number of patients in which skin tests became negative after the procedure. To assess safety, all breakthrough reactions were collected.


Flow of assessed reactions through anamnesis, skin testing, risk assessment, graded challenge

During a 1-year period, 58 patients were referred, and 23 of them met the inclusion criteria to be treated with rapid desensitization (Fig. 1).

Figure 1.

Flow diagram of assessed reactions in all patients from first referral to inclusion/exclusion for rapid desensitization. Divided into platins, taxanes, and other drugs. There are 24 initial reactions amenable to undergoing rapid desensitization in 23 candidate patients because one patient met the inclusion criteria for 2 different drugs (docetaxel and cyclophosphamide). Excluded reactions are included in crossed boxes (image).

Characteristics of candidate patients for desensitization

For the group of 23 desensitized patients, mean age was 56 years old, and 13 of them (56.5%) were women. All patients were being treated for malignancy, mainly breast cancer (9/23), and colorectal cancer (8/23); but also for pancreatic cancer (2/23), non-Hodgkin lymphoma (2/23), lung cancer (1/23), and ovarian cancer (1/23). The prevalence of atopy (allergic rhinitis, allergic conjunctivitis, allergic asthma, atopic dermatitis, or food allergy) was 30.4%. The prevalence of confirmed drug allergy was 13%. Culprit drugs in initial reactions (to which patients were afterward desensitized) were platins (11/23; being 10/23 oxaliplatin and 1/23 carboplatin), taxanes (10/23; being 9/23 paclitaxel and 1/23 docetaxel), cyclophosphamide (1/23), and rituximab (2/23). One patient was desensitized to two different drugs (cyclophosphamide and docetaxel).

In the group of 23 desensitized patients, initial reaction was moderate or severe (grade 2–4) in 65% of patients, and only 35% (8/23) of patients could be classified as having suffered a mild initial reaction (grade 1). Characteristics of initial reactions are shown in Fig. 2.

Figure 2.

Characteristics and severity of initial reactions and breakthrough reactions.

Desensitization outcomes

Of 189 intravenous rapid desensitizations undertaken, 188 were successfully accomplished with the RCUH 10-step protocol (38 oxaliplatin, 5 carboplatin, 137 paclitaxel, 2 docetaxel, 1 cyclophosphamide, 5 rituximab). Of 23 patients, 22 could receive their target dose (one patient treated with paclitaxel revoked her consent to finish the procedure after breakthrough reaction).

No reactions occurred in 94% of desensitizations (177/189), and mild reactions (grade 1) were observed in 4% of desensitizations (8/189). Moderate or severe reactions were only observed in 4 desensitizations. No reactions graded 4 (anaphylaxis) or 5 (death) were observed. A total of 13 reactions were observed in 11 patients (48% of all desensitized patients; being 6/11 oxaliplatin-reactive patients and 5/11 paclitaxel-reactive patients), with 1 paclitaxel-reactive patient suffering 2 reactions during the same desensitization. Further data on breakthrough reactions (compared with initial reactions) are shown in Fig. 2.

Of the total of 13 breakthrough reactions, 2 reactions (15%) occurred during the infusion of solution A (steps 1–4). During the infusion of solution B (steps 5–8), we also observed 2 reactions; and, all the other reactions (69%) were observed during the infusion of solution C (steps 9–10). All reactive patients suffered their breakthrough reactions during the first desensitization, except for one oxaliplatin-reactive patient who suffered another reaction during the second desensitization, and needed additional steps.

Skin testing was undergone immediately after successful desensitizations, in all patients under treatment with platins and previous positive skin tests, showing negative results in 100% of patients. This could not be confirmed for taxanes because histamine control was inhibited (taxanes standard premedications included antihistamines and corticosteroids).

Oxaliplatin-specific IgE assessment

During the 1-year period of the study, 23 oxaliplatin-reactive patients were referred to the program. The results of their study with skin testing, controlled challenge, and ImmunoCAP® are shown in Table 4.

Table 4. Results of the study with skin testing, controlled challenge, and ImmunoCAP® in patients reacting with oxaliplatin
GenderAgePrick 5 mg/mlID 0.5 mg/mlID 5 mg/mlControlled ChallengeSpecific IgE (UI/l)
ImmunoCAP® assessment as a diagnostic tool
 Cutoff point 0.10UI/l (%)Cutoff point 0.35UI/l (%)
  1. Neg, negative; Pos, positive; F, female; M, male; NC, patient does not sign informed consent; ID, intradermal test; positive results for ImmunoCAP® are enhanced in gray shade.

  2. For the ImmunoCAP® assessment as a diagnostic tool, of the 23 oxaliplatin-reactive patients, only 16 patients with either positive or negative allergological study (Gold Standard) were included.



A total of 189 desensitizations to several antineoplastic agents were undergone, by means of a new, short, rapid desensitization protocol in a 1-year period. All patients except for one (who revoked her consent to finish the procedure) could receive their target dose, proving this protocol effective in administering antineoplastic agents to patients who had previously suffered HSR to them.

Most desensitizations (94%) were accomplished with no breakthrough reactions, and most reactions (61.5%) were mild. Moderate or severe reactions were only observed during 4 desensitizations. No deaths or anaphylaxis were reported. Most breakthrough reactions occurred during the last steps of the protocol (69% during steps 9–10), as it has been observed by other groups [1]. These results show no obvious increase in the number or severity of breakthrough reactions when compared to results reported by other groups [1]. However, different safety profiles could be influenced by differences in populations.

Most desensitizations were achieved with no breakthrough reactions, but approximately half of the desensitized patients (11/23) suffered a reaction during their first desensitization. The use of premedication with montelukast and acetylsalicylic acid (protocol unaltered) made the protocol successfully tolerated for all patients who went on with a second desensitization, except for one oxaliplatin-reactive patient who needed additional steps. The use of 500 mg of acetylsalicylic acid instead of 325 mg did not cause adverse effects on any patients. We think that systematic premedication with acetylsalicylic acid and montelukast may improve desensitization tolerability and could be implemented to reduce the number of reactive patients, and this is coherent with previous data [8].

Our population of desensitized patients shows a higher representation of male patients than other series [14]. Initial reactions were mainly moderate to severe, with only a 35% of patients having suffered a mild initial reaction. Atopy and drug allergy prevalence were not clearly higher in this population than in general population [25]. Culprit drugs also differ from other series, reporting 10 oxaliplatin-reactive patients who underwent 38 desensitizations, with a total of 6 breakthrough reactions in 5 oxaliplatin-desensitized patients (all mild, except for one moderate reaction).

Previous experience [22, 26] finds skin testing with taxanes of little diagnostic value; however, we found two patients with positive skin tests to paclitaxel. Moreover, cases of IgE-dependent reactions to paclitaxel have been reported [27]. We think skin testing with taxanes could be useful as a risk marker.

We report the first published data, up to our knowledge, on oxaliplatin-specific IgE determination by ImmunoCAP®: In our selected population, this diagnostic test is very specific but less sensitive. Selection bias may overestimate specificity and underestimate sensitivity. More studies are necessary to validate these diagnostic tools.

Sixty percent of patients referred to the program (35/58) had the opportunity to continue with their therapy (either by the use of desensitization or after a negative-graded challenge) within the Desensitization Program; however, 22 patients changed their treatments due to oncology indication or delayed reactions (Fig. 1).

In conclusion, in the hands of a Desensitization Program, this new desensitization protocol has proven an effective and safe therapeutic tool for hypersensitivity to antineoplastic agents. It is designed to last approximately 4 h and permits single-morning desensitizations. It was modified from the previously in vivo and in vitro standardized BWH protocol, to comply with safety measures for hazardous drugs handling at many institutions. It has been successfully used for different drugs (oxaliplatin, carboplatin, paclitaxel, docetaxel, cyclophosphamide, and rituximab) and for different types of immediate reactions (IgE-dependent and IgE-independent). We present the first important series of oxaliplatin desensitizations to our knowledge; and, additionally, we find oxaliplatin-specific IgE determination may be a helpful novel diagnostic tool. Finally, we think universal establishment of Desensitization Programs to ensure patient's safe first-choice therapies, under multidisciplinary team approach by drug desensitization experts and specialty trained personnel in appropriate installations, should be a standard of care in the 21st century.


We want to thank all the personnel involved in the RCUH Desensitization Program and ThermoFisher Scientific Phadia AB. We want to thank Mariana Castells and the BWH Desensitization Program team, for their inspirational efforts in research and development of desensitization techniques, which bring hope for many patients.

Author contributions

All the authors have contributed to the design of the article, and/or to the interpretation of data, and/or to drafting and critically revising the article for important intellectual content, and/or to the final approval of the version to be published.

Conflicts of interest

There are no potential conflicts of interest for any of the authors regarding this article. There are no financial interests, and there has not been any provision of study materials by their manufacturer for free or at a discount from current rates.