Clinical value of in vitro tests for the management of severe drug hypersensitivity reactions.

Drug hypersensitivity reactions (DHRs) occasionally present with severe cutaneous adverse reactions (SCARs) which result in a high risk of morbidity and mortality. Although SCARs are rare, the occurrence could lead to a significant increase in healthcare and economic burden, especially when more than one possible culprit drug is implicated. Therefore, the accurate identification of the culprit drug(s) is important for correct labeling and subsequent patient education and avoidance. To date, clinical evaluation using causality assessment has limitations because the assessment may be inaccurate due to the overlapping timelines when multiple drugs are initiated/continued. Moreover, drug provocation tests (DPTs) which is the gold standard in diagnosis, are contraindicated, and in vivo skin tests may also be associated with risks of triggering SCAR. The European Network for Drug Allergy recommended that in vitro tests, if available, should be performed before any in vivo tests. Basophil activation tests and lymphocyte transformation tests, could serve as reliable in vitro tests for both immediate and delayed-type DHR. Many academic medical centers with affiliated laboratory services offer these tests in the diagnostic evaluation of SCARs in clinical practice. This not only complements identification of the culprit drug(s), but may also be used to test for potentially non cross-reactive alternatives, hence avoiding DPTs. In this review, we summarize the roles of in vitro tests in identifying the culprit drug(s) in SCARs, issues with utilization and interpretation of test results, and our experience in clinical practice.

INTRODUCTION

Drug hypersensitivity reactions (DHRs) are an important public health problem. DHRs manifest with mild to life-threatening symptoms in susceptible individuals [1]. Life-threatening manifestations of DHR in the form of severe cutaneous adverse reactions (SCARs), are associated with high risks of morbidity and mortality, albeit its low prevalence [1]. Risk factors for SCARs include host genetic factors, threshold/eliciting dose of the drug, renal impairment, and inappropriate immune responses. Therefore, diagnosis of SCARs involves the correlation between timeline of symptoms, and corroborative clinical and laboratory findings in relation to the history of drug exposure [23]. Finally, a definitive diagnosis with causative drug identification is needed. To start to identify culprit drug causing SCARs, one could gather a history of drug exposure in detail as well as a timeline of SCAR onset and progression; to include the possible culprit drugs and exclude temporally irrelevant drugs [2]. However, this causality assessment has limitations because the outcome is usually inaccurate due to the overlapping timelines [4].

Skin tests (STs) are usually performed by well-trained allergists or dermatologists. However, the diagnostic utility of these tests varies with the agents used, concentration, lack of standardization of ST preparations, each with an unknown predictive value. Although STs have been evaluated as having acceptable sensitivity to identify some common culprit drugs, such as beta-lactam, perioperative drugs, radiocontrast media (RCM), and platinum-based chemotherapeutic drug, they still have low diagnostic potency for most drugs. In addition, validation of ST protocols are lacking, and their sensivity is suboptimal [5]. To identify culprit drug causing immediate drug hypersensitivity, skin prick test (SPT) and intradermal test (IDT) have been widely used as both tests could provide high sensitivity and very high negative predictive value, especially in the case of beta-lactam allergy [67]. However, SPT is relatively contraindicated in pregnant people; and IDT is contraindicated in SCAR as it has a high risk of causing a systemic reaction [68]. Additionally, it is difficult to perform ST in patients with severe dermatitis, dermographism or taking immunomodulating or antihistamine drugs [6]. For delayed-type drug hypersensitivity, delayed-reading IDT, and patch test (PT) have been employed to identify culprit drugs. Both IDT and PT provided high negative predictive value, however, IDT has higher sensitivity than PT. As IDT is contraindicated in severe drug reactions, hence PT has been safer and more widely performed with any form of drugs [8]. Positive rates of PT depend on type of SCARs, drug, and vehicle [7].

DPT has been considered to be a gold standard for DHR diagnosis if the observed reactions are compatible with drug allergy. Nevertheless, DPT is contraindicated in cases with SCAR. The position paper from the European Academy of Allergy and Clinical Immunology/European Network of Drug Allergy recommended that if the in vitro tests are available, they could be the first choice to help the clinician to arrive at the correct diagnosis [9]. These tests do not only identify culprit drugs with excellent specificity, but they also do not put vulnerable patients under a high-risk procedure. Although the in vitro tests often provide poor sensitivity, they could complement identification of the culprit drug, and assist physicians to seek safer alternative drugs.

The in vitro tests may be considered before in vivo procedures. Ideally, the diagnostic tests with high accuracy (high sensitivity and specificity) should be useful for high-risk patients with severe index reactions. A positive test result could provide a firm diagnosis of the culprit drug, therefore, decreasing the need for in vivo or unnecessary DPTs. A negative test result could decrease the likelihood of certain drugs being the putative drug, the physicians could move on to a further in vivo test or sometimes DPTs [10]. Hence, the in vitro tests, if available, could be considered as the investigation of choice in high-risk patients. However, there are many concerning issues involving the management of in vitro tests that will be discussed in this review.

The objective of this article is to review clinical applications of the in vitro tests for immediate-type DHR (basophil activation test, BAT) and delayed-type DHR (lymphocyte transformation test, LTT) for which we recently published our experience from Thailand [11121314]. Some concerning issues for the in vitro test interpretation and laboratory limitations are also included. We summarize the utility of in vitro tests, laboratory interpretation, and clinical interpretation to physicians, not only to improve the understanding of these tests, but also to increase awareness of the limitations of the in vitro tests.

IMPLICATION OF IN VITRO TESTS IN CLINICAL PRACTICE

The in vitro tests have been developed according to the types of DHRs. So far, there have been many in vitro tests developed for culprit drug identification [9]. Basophil histamine release (BHR) assay and BAT have been established for immediate-type DHRs. However, the latter has been used more widely because BAT provides significantly higher sensitivity as compared to BHR. BAT can be performed with low number of basophils, while BHR cannot [1516]. For delayed-type DHR, the in vitro tests have been developed, such as cytokine detection by ELISA or ELISpot and LTT [17]. Although ELISA and ELISpot are moderately complicated tests and require a few days to complete the whole assay, standardization of cutoff values, selection of suitable cytokines for various types of delayed-type DHR remain issues. Even LTT is a laborious assay and needs an incubation time for cell proliferaton, it has been reliably utilized as a tool to identify culprit drug for delayed-type DHR for T-cell proliferation[18]. Therefore, the scope of this review will focus on BAT and LTT.

BAT for immediate drug hypersensitivity

The BAT is a cell-based assay using flow cytometry to evaluate the status of basophils by measuring changes in activation markers (commonly CD63 and/or CD203c) expressed on the cell membrane of basophils [15]. The general principle of BAT is that CD63, which is a membrane-bound molecule and intragranularly expressed, would be exposed on the cell surface during degranulation where granular membrane fuses with the cellular membrane. Additionally, CD203c, which is constitutively expressed on the basophil surface membrane, is significantly upregulated when basophils are activated. Even though the principles behind BATs are straightforward, the use of multiple negative and positive controls as well as consistent cytometer settings are required in order to limit the risk of incorrect reporting and interpretation of results.

BAT has been shown to be a reliable tool in culprit drug identification even though wide sensitivity (28%–100%) and specificity (80%–100%) have been reported. Examples of drugs commonly identified with BAT are RCM, beta-lactams, quinolones, nonsteroidal anti-inflammatory drugs (NSAIDs), and neuromuscular blocking agents (NMBAs) as summarized in Table 1 [1920212223242526272829]. This table also shows the current literature published on the value of BAT in the diagnosis of immediate DHR to beta-lactams, quinolones, NMBAs, RCM, and NSAIDs. Additionally, our in-house BAT studies (unpublished data) on these drugs and antituberculosis drugs were included. It appears that CD63 was consistently used as a primary marker for basophil activation and CD203c was employed as a complementary marker. It seemed that cutoff point was quite consistent criteria when percentage of basophil positive for activation markers was ≥5% and stimulation index (SI as calculated by a ratio of percentage of basophil positive between drug-stimulated condition and unstimulated condition) was ≥2 (according to manufacturer’s recommendation). In our studies using in-house BAT, both CD63, and CD203c were evaluated and BAT was interpreted as positive when either CD63 or CD203c reached cutoff criteria (≥5% and SI≥2). We included both activation markers as different drugs could possibly stimulate different mechanisms leading to expression of different activation markers. Additionally, we also successfully developed BAT to identify chlorhexidine and ortho-phthalaldehyde as culprit drugs in anaphylactic cases [1330]. In conclusion, although sensitivity of BAT varied among these studies, its specificity was excellent.

Table 1

Utility of BAT for culprit drug identification in immediate DHR

Drugs		Markers	Cutoff	Sensitivity (%)	Specificity (%)	Reference
Beta-lactams
	AMX, AMP, PEN, CEFU	CD63	≥5%, SI≥2	49–55	91–100	[1920]
	AMX, AMP, CEF-3, CEFT, CEFU	CD63, CD203c	≥5%, SI≥2	33	100	*
Quinolones
	MOX, LEV, CIP, OLF	CD63, CD203c	≥5%, SI≥2	71	100	[2122]
NMBAs
	PAN, ROC, ATA, SUX	CD63, CD203c	>4%, >10%	80–100	96–100	[2324]
RCM
	IOB, IOX, IOP, IOH, IOPA, IOM	CD63	≥5%, SI≥2	46–63	89–100	[2526]
	IOB, IOH, IOP, IOX	CD63, CD203c	≥5%, SI≥2	25	ND	*
NSAIDs
	ASA, DIC, KET, CEL, ACT	CD63	≥5%, SI≥ (1.71–2.18)	37–61	90–91	[272829]
	DIC, IBU, ACT	CD63, CD203c	≥5%, SI≥2	33	100	*

BAT, basophil activation tests; DHR, drug hypersensitivity reactions NMBAs, neuromuscular blocking agents; RCM, radiocontrast media; NSAIDs, nonsteroidal anti-inflammatory drugs containing ibuprofen, diclofenac, tramadol; ND, not determined; AMX, amoxicillin; AMP, ampicillin; PEN, penicillin; CEFU, cefuroxime; CEF-3, ceftriaxone; CEFT, ceftazidime; MOX, moxifloxacin; LEV, levofloxacin; CIP, ciprofloxacin; OLF, ofloxacin; PAN, pancuronium; ROC, rocuronium; ATA, atacurium; SUX, sumamethonium; IOB, iobitridol; IOH, iohexol; IOM, iomeprol; IOP, iopromide; IOPA, iopamidol; IOX, ioxithalamate; ASA, aspirin; DIC, diclofenac; KET, ketoprofen; CEL, celecoxib; ACT, acetaminophen; SI, stimulation index.

*Data of our studies, unpublished data.

Discussion points for BAT setting

Interestingly, basophils from 5%–20% of patients do not respond to BAT positive controls, therefore, it is hard to interpret results for diagnosis. To increase the chance of positive control responsiveness, our in-house setting for BAT included various positive controls for basophil activation, IgE-dependent (anti-IgE and anti-FceRI), and IgE-independent (N-formyl-L-methionyl-L-leucyl-phenylalanine), as a positive control panel. In addition, a wide range of drug concentrations (at least 3 concentrations) were set in our assay. Additionally, a negative control consisting of stimulation buffer alone should also be included to assess the level of background or spontaneous activation of basophils. In case of interleukin (IL)-3 priming, an additional negative control for IL-3 in the stimulation buffer should be included. It is important to set standardization, quality assurance of the laboratory procedures, flow cytometry acquisition, and data analyses in order for laboratory BAT to be applied correctly to clinical practice. In addition, clinical validation with BAT studies could be performed to assure the accuracy of test results. As mentioned earlier, a wide range of BAT sensitivity was reported, which could be due to (1) nonestablished standardization of BAT (various methods, techniques and instrument setting), (2) type and optimal concentration of drug, (3) cutoff criteria, (4) quality of fresh blood sample, and (5) recruitment of appropriate cases with well-defined clinical histories [3132].

BAT is not useful for evaluation of nonallergic hypersensitivity to any stimulants if the stimulants can activate via alternative IgE-independent effector cell activation mechanisms, such as through off-target occupation of the MRGPRX2 receptor [33]. This event can occur in patients who have positive ST but negative BAT because skin mast cells express MRGPRX2, unlike basophils which barely express such receptor. Therefore, BAT may probably give low sensitivity in case of MRGPRX2 activation leading to DHR in the case of NSAIDs and quinolone hypersensitivity [3134].

LTT for delayed drug hypersensitivity

LTT is a cellular testing approach to reveal the existence of drug-specific memory lymphocytes in circulating blood and its proliferative response to identify culprit drug and confirm DHR. These drug-specific lymphocyte responses rely on 3H-thymidine incorporation during lymphocyte proliferation. Its specificity was 63%–100% (mostly >90%), whereas its sensitivity varied depending on the drug and types of DHR as summarized in Table 2 [1718353637383940414243].

Table 2

Utility of LTT for culprit drug identification in delayed-type DHR

Drugs		Clinical manifestation	Sensitivity (%)	Specificity (%)	Reference
Beta-lactams
	AMX, AMP, PEN, BEN	MPE, EXT, AGEP, TEN	68–83	85–100	[1835363738]
	AMX, AMP, PEN, CEF-3	MPE, DRESS, AGEP, SJS/TEN	21	100	*
Anti-TB drugs
	IRZE	BUL, EXT, DILI	29–87	90–100	[394041]
	IRZEL	MPE, DRESS, SJS/TEN	52	89	*
Antibiotics
	VAN, COT, CIP, LEV, OLF, MOX, MER, CLO, PIP	LABD, DRESS, AGEP, SJS/TEN	77	100	*
Antiepileptic drugs
	CBZ, LTG	MPE, EXT, SJS/TEN	26–66	63–100	[17374243]
	CBZ, PHE, VAL, LTG	DRESS, SJS/TEN	55	91	[11], *
	NSAIDs
	FEN, FLU	TEN	44	63	[43]
	IBU, TRA, ACT, DIC, MEF	DRESS, SJS/TEN	50	98	*

LTT, lymphocyte transformation tests; DHR, drug hypersensitivity reaction; AMX, amoxicillin; AMP, ampicillin; PEN, penicillin; BEN, benzylpenicillin; CEF-3, ceftriaxone; MPE, maculopapular eruption; EXT, exanthem; DRESS, drug reaction with eosinophilia and systemic symptoms; SJS/TEN, Stevens-Johnson syndrome/toxic epidermal necrolysis; AGEP, acute generalised exanthematous pustulosis; TB, tuberculosis; IRZEL, isoniazid, rifampicin, pyrazinamide, ethambutol, levofloxacin; BUL, bullous; DILI, drug-induced liver injury; NSAIDs, nonsteroidal anti-inflammatory drugs containing ibuprofen, diclofenac, mefenamic acid, acetaminophen-tramadol; VAN, vancomycin; COT, cotrimoxazole; LEV, levofloxacin; CIP, ciprofloxacin; OLF, ofloxacin; MOX, moxifloxacin; MER, meropenem; CLO, cloxacillin; PIP, piperacillin; LABD, linear immunoglobulin A bullous dermatosis; CBZ, carbamazepine; PHE, phenytoin; VAL, valproic acid; LTG, lamotrigine; FEN, fenbrufen; FLU, flurbiprofen; IBU, ibuprofen; TRA, tramadol; DIC, diclofenac; ACT, acetaminophen; MEF, mefenamic acid.

*Data of our studies, unpublished data.

LTT has been used as a tool to identify culprit drug for diagnosis of delayed DHR. As summarized in Table 2, the most common culprit drugs reported as culprit drugs identified by LTT were beta-lactams and antiepileptic drugs. Additionally, LTT could identify causative drugs for delayed DHR in other drug groups such as anti-TB drugs, cotrimoxazole, quinolones, vancomycin, and NSAIDs. LTT studies for identification of anti-TB drugs as culprit drugs have been reported from Asian countries. Our LTT setting has also been used to identify culprit drugs from various drugs as aforementioned. In addition, other drugs, such as allopurinol, antiviral drug, and heavy metal (nickel) have also been identified by our LTT setting. In our experience, positive results of LTT correlated with well-defined cases (consistent history, timeline, and clinical manifestation) and high score of Naranjo’s assessment [11]. According to many reports, LTT yielded higher sensitivity than STs, making it a promising in vitro diagnostic tool for delayed type DHR [183544].

Discussion points for LTT setting

To set LTT standardization, quality assurance of the cell culture procedures, such as capacity for cell growth of each preparation batch of culture medium and supplements (especially pooled AB plasma) and capacity to induce cell proliferation of each lot of positive control (phytohemagglutinin [PHA] and tetanus toxoid), have to be regularly assessed [45]. In order to interpret LTT results, sufficient amount and percentage of viable peripheral blood mononuclear cells (PBMCs) were routinely evaluated in our LTT setting. Positive controls in our LTT included nonspecific mitogen (PHA) and recall antigens (tetanus toxoid and purified protein derivative (in the case of tuberculosis). Every concentration of drugs or reagents was also tested as nontoxic levels for cell culture. In our LTT setting, irrelevant drugs were also included to clarify specificity of the test. Nevertheless, false-positive results could occur when PBMCs were stimulated with some drugs, such as vancomycin, acetaminophen and RCMs [45] and valproic acid (our finding; unpublished data), or when LTT was performed at an inappropriate time (such as acute phase in DRESS [drug reaction, eosinophilia with systemic symptoms]) [46]. In vitro tests are usually performed after remission of disease, as PBMCs obtained during the acute drug allergy stage are strongly activated due to bystander activation of innate cells. High background proliferation might make it difficult to detect proliferation of drug-specific T-cell clones. The time interval between the acute stage and test performance has been recommended as 4–8 weeks after recovery of delayed-type hypersensitivity reaction and extended for up to 12 weeks after DRESS remission to wash out the incriminated drug and any other drugs [4647]. Even though LTT could provide very high specificity (85%–100%), wide range of sensitivity could occur according to (1) recruitment of cases with mixed-up Naranjo’s score and not well-defined clinical history, (2) drugs, (3) threshold of cutoff point, and (4) clinical manifestations (maculopapular eruption, Stevens-Johnson syndrome/toxic epidermal necrolysis) [45]. We have previously demonstrated that SI well corresponded to Naranjo’s score and SI ≥ 2.0 could be the optimal cutoff point after receiver operating characteristic analysis when compared with other SI values [11].

LABORATORY RESULTS CONTRIBUTE TO CLINICAL PRACTICE

Positive results

As these in vitro tests could provide excellent specificity, positive results could identify culprit drug and provide definite DHR diagnosis. The positive results could convince allergy practitioners that the suspected drug could be the cause of DHR and no further in vivo investigation is needed.

Negative results

Although negative results from these tests with varying sensitivity could not rule out a suspected drug, they maintain sufficient negative predictive value for allergists to work up with some clues represented by in vitro nonresponsiveness. With the nonresponsive clues, they may help the allergist manage and ensure availability of appropriate drugs for further in vivo test or drug provocation test. Some representative cases of our experience to perform in vivo test or challenge test according to clues from in vitro tests are demonstrated in Table 3.

Table 3

Utilization of in vitro tests for severe DHRs in our clinical practice

Immediate DHR
	DHRs manifestation	Possible culprits	In vitro-tested drugs	BAT	STs	DPT
	Anaphylaxis	Lidocaine	Lidocaine	Negative	Negative	ND
			Mepivacaine	Negative	Negative	Negative
		Final drug use: Mepivacaine
	Anaphylaxis	Iobitridol	Iobitridol	Negative	ND	ND
			Iopromide	Negative	ND	Negative
			Iohexol	Negative	ND	ND
			Ioxagate	Negative	ND	ND
		Final drug use: Iopromide (no premed, no ADR)
	Possible Kounis syndrome	Amoxicillin/clavulanic	Amoxicillin	Negative	Negative	ND
			Clavulanic	Negative	Negative	ND
		Continue to avoid amoxicillin/clavulanic
Delayed-type DHR
	DHRs manifestation	Possible culprits	In vitro-tested drugs	LTT	STs	DPT
	DRESS	IRZE	Isoniazid (I)	Negative	ND	Negative
			Rifampicin (R)	Positive	ND	ND
			Ethambutol (E)	Negative	ND	Negative
			Pyrazinamide (Z)	Positive	ND	ND
			Levofloxacin (L)	Negative	ND	Negative
		Final regimen: IEL
	DRESS	IRZE	Isoniazid (I)	Negative	ND	Negative
			Rifampicin (R)	Negative	ND	Negative
			Ethambutol (E)	Negative	ND	Negative
			Pyrazinamide (Z)	Positive	ND	ND
			Levofloxacin (L)	Positive	ND	ND
		Final regimen: IRE

DHR, drug hypersensitivity reaction; BAT, basophil activation test; ST, skin test; DPT, drug provocation test; ND, not determined; ADR, adverse drug reaction; DRESS, drug reaction with eosinophilia and systemic symptoms; LTT, lymphocyte transformation test; IRZE, isoniazid, rifampicin, pyrazinamide, ethambutol.

According to our finding in Table 3, the majority of negative in vitro cases could pass a drug provocation/challenge test or be confirmed by in vivo tests. This suggests that patients may not really be allergic to the suspected drug or have a chance to use alternative drugs. Our management was consistent with other studies showing that DHR patients, who obtained negative in vitro test with unresponsiveness in challenge test or DPT, had a chance to successfully retake the suspected drugs [10].

CONCERNED INTERPRETATION

As in vivo and in vitro tests for DHR provide low-to-moderate sensitivity; technical improvement and test combination are required to enhance sensitivity of these tests [48]. Positive results from tests with excellent specificity and negative predictive value could be useful to provide definitive diagnosis of DHR with culprit drug identified. However, negative results from low-to-moderate sensitivity and positive predictive value tests cannot be used to rule a suspected drug out, and further investigation or avoidance of the suspected drug is required. Likewise, false-positive and false-negative results have to be considered and analyzed during results interpretation. False-negative results may lead to serious outcomes, which can contribute to patient safety with re-exposure of culprit drug. Meanwhile, false-positive results do not affect patient safety, however, they make patients lose a chance to receive the first-line drugs with high potency. Nevertheless, investigation for culprit drug identification is necessary to give a definite diagnosis of DHR as well as no or very low risk from investigation has to be concerned. With such a situation, therefore, in vitro tests could be the most appropriate investigation to provide evidence for physicians to plan further treatment. Many factors contributing to either false-negative or false-positive results were summarized in Table 4.

Table 4

Factors affecting results of BAT and LTT

Outcomes	BAT	LTT
False-negative results may be obtained if [313245]	Performed within 1–2 weeks after recently reaction (refractory period)	-
	Systemic corticosteroid, immunosuppressant drugs (cyclosporin A)	Systemic corticosteroid, immunosuppressant drugs, chemotherapeutic drugs
	Last reaction > 1 yr	Last reaction > 2–3 yr
	Prolonged blood handling > 24 hr	Prolonged blood handling > 24 hr
	Nonresponder to controls	Nonresponder to controls
	-	High/toxic doses lead to cell death
False-positive results may be obtained if [3245]	-	AGEP, SJS/TEN: within 1 month after recovery
		DRESS: within 3 months after recovery
	-	Some drugs (vancomycin, acetaminophen, RCM, NSAIDs)
	High/irritating doses leading to nonspecific degranulation	-

BAT, basophil activation test; LTT, lymphocyte transformation test; AGEP, acute generalised exanthematous pustulosis; DRESS, drug reaction with eosinophilia and systemic symptoms; SJS/TEN, Stevens-Johnson syndrome/toxic epidermal necrolysis; RCM, radiocontrast media; NSAIDs, nonsteroidal anti-inflammatory drugs.

Of note, antihistamine has no effect on BAT (both CD63 and CD203c), which make BAT more superior than in vivo test in patients who could not omit antihistamine [324950]. LTT could be suitable in some decent conditions which may be performed with blood samples drawn from patients who take <0.2 mg/kg/day of prednisolone or who take methotrexate/azathioprine without lymphopenia [45]. Even though there was no consensus on standardization of ELISpot to identify culprit drug in delayed-type DHR, in case culprit drug identification is urgently needed, IFN-γ ELISpot might be an alternative assay as reported [12].

CONCLUSION AND PERSPECTIVE

In vitro tests are promising tools that can help allergists to identify culprit drugs for the definitive diagnosis of DHR. In vitro tests have been shown to be useful when in vivo diagnosis are lacking or not possible. To evaluate the role of in vitro tests in clinical practice for DHR diagnosis, they require DPT outcomes in many cases for validation analysis. However, with ethical reasons, DPTs cannot be performed in many patients with life-threatening reactions to serve as test validation. With this situation, clinical data with thorough assessment and STs, if available, are employed.

As a result of the prevalence of DHRs, another limitation of many studies is the small sample size. Therefore it is likely that multicenter studies will be needed because these can facilitate the harmonization of techniques and inclusion of sufficient numbers of patients/control subjects.