Designing Predictive Models for Beta-Lactam Allergy Using the Drug Allergy and Hypersensitivity Database

Predictive models and applicability of artificial intelligence-based approaches in drug allergy

PURPOSE OF REVIEW

Drug allergy is responsible for a huge burden on public healthcare systems, representing in some instances a threat for patient's life. Diagnosis is complex due to the heterogeneity of clinical phenotypes and mechanisms involved, the limitations of in vitro tests, and the associated risk to in vivo tests. Predictive models, including those using recent advances in artificial intelligence, may circumvent these drawbacks, leading to an appropriate classification of patients and improving their management in clinical settings.

RECENT FINDINGS

Scores and predictive models to assess drug allergy development, including patient risk stratification, are scarce and usually apply logistic regression analysis. Over recent years, different methods encompassed under the general umbrella of artificial intelligence, including machine and deep learning, and artificial neural networks, are emerging as powerful tools to provide reliable and optimal models for clinical diagnosis, prediction, and precision medicine in different types of drug allergy.

SUMMARY

This review provides general concepts and current evidence supporting the potential utility of predictive models and artificial intelligence branches in drug allergy diagnosis.

Risk factors for drug hypersensitivity reactions in children

Drug hypersensitivity reactions are common in children. Risk factors predisposing to IgE-mediated drug allergies and delayed drug reactions are a matter of debate. Gender, age, previous reactions to the same drug or to another drug, reduced drug metabolism, chronic diseases, polypharmacy, drug doses are linked with the onset of hypersensitivity reactions in some children. Novel advances in genetic polymorphisms can rapidly change the approach to the prevention of reactions since gene testing can be a useful screening test for severe cutaneous adverse reactions. Viral infections may act as cofactors in susceptible individuals. Polypharmacy, high doses, repeated doses and parental route of administration are also risk factors. Clinicians should take into account risk factors to allow the risk-benefit balance to be maintained.

Predicting Penicillin Allergy: A United States Multicenter Retrospective Study

BACKGROUND

Using the reaction history in logistic regression and machine learning (ML) models to predict penicillin allergy has been reported based on non-US data.

OBJECTIVE

We developed ML positive penicillin allergy testing prediction models from multisite US data.

METHODS

Retrospective data from 4 US-based hospitals were grouped into 4 datasets: enriched training (1:3 case-control matched cohort), enriched testing, nonenriched internal testing, and nonenriched external testing. ML algorithms were used for model development. We determined area under the curve (AUC) and applied the Shapley Additive exPlanations (SHAP) framework to interpret risk drivers.

RESULTS

Of 4777 patients (mean age 60 [standard deviation: 17] years; 68% women, 91% White, and 86% non-Hispanic) evaluated for penicillin allergy labels, 513 (11%) had positive penicillin allergy testing. Model input variables were frequently missing: immediate or delayed onset (71%), signs or symptoms (13%), and treatment (31%). The gradient-boosted model was the strongest model with an AUC of 0.67 (95% confidence interval [CI]: 0.57-0.77), which improved to 0.87 (95% CI: 0.73-1) when only cases with complete data were used. Top SHAP drivers for positive testing were reactions within the last year and reactions requiring medical attention; female sex and reaction of hives/urticaria were also positive drivers.

CONCLUSIONS

An ML prediction model for positive penicillin allergy skin testing using US-based retrospective data did not achieve performance strong enough for acceptance and adoption. The optimal ML prediction model for positive penicillin allergy testing was driven by time since reaction, seek medical attention, female sex, and hives/urticaria.

How to Define and Manage Low-Risk Drug Allergy Labels

Risk stratification in drug allergy implies that specific risk categories (eg, low, moderate, and high) classify historical drug hypersensitivity reactions. These risk categories can be based on reaction phenotypic characteristics, the timing of the reaction and evaluation, the required reaction management, and individual characteristics. Although a multitude of frameworks have been described in the literature, particularly for penicillin allergy labels, there has yet to be a global consensus, and approaches continue to vary between allergy centers. Immune-mediated drug allergies can sometimes be confirmed using skin testing, but a negative drug challenge is required to demonstrate tolerance and remove the allergy from the electronic health record ("delabel" the allergy). Even for quintessential IgE-mediated drug allergy, penicillin allergy, recent data reveal that a direct oral challenge, without prior skin testing, is an appropriate diagnostic strategy in those who are considered low-risk. Drug allergy pathogenesis and clinical manifestations may vary depending on the culprit drug, and as such, the optimal approach should be based on risk stratification that considers individual patient and reaction characteristics, the likely hypersensitivity reaction phenotype, the drug class, and the patient's clinical needs. This article will describe low-risk drug allergy labels, focusing on β-lactam and sulfonamide antibiotics, nonsteroidal anti-inflammatory drugs, iodinated contrast media, and common chemotherapeutics. This review will also address practical management approaches using currently available risk stratification and clinical decision tools.

Delabeling of allergy to beta-lactam antibiotics in hospitalized patients: a prospective study evaluating cost savings

BACKGROUND

Patients with a penicillin allergy label are at risk of an associated increase in adverse antibiotic events and hospitalization costs.

AIM

We aimed to study the economic savings derived from the correct diagnosis and delabeling inpatients with suspected beta-lactam allergy, considering the acquisition cost of antimicrobials prescribed during a patient's hospital stay.

METHOD

We prospectively evaluated patients admitted to the University Hospital of Salamanca who had been labeled as allergic to beta-lactams and performed a delabeling study. Subsequently, cost differences between antibiotics administered before and after the allergy study and those derived from those patients who received alternative antibiotics during admission and those who switched to beta-lactams after the allergy study were calculated.

RESULTS

One hundred seventy-seven inpatients labeled as allergic to beta-lactams underwent a delabeling study; 34 (19.2%) were confirmed to have allergy to beta-lactams. Of the total number of patients, 136 (76.8%) received antibiotics during their hospitalization, involving a mean (SD) cost of €203.07 (318.42) and a median (IQR) cost of €88.97 (48.86-233.56). After delabeling in 85 (62.5%) patients, the antibiotic treatment was changed to beta-lactams. In this group of patients, the mean cost (SD) decreased from €188.91 (351.09) before the change to 91.31 (136.07) afterward, and the median cost (IQR) decreased from €72.92 (45.82-211.99) to €19.24 (11.66-168). The reduction was significant compared to the median cost of patients whose treatment was not changed to beta-lactams (p<0.001).

CONCLUSION

Delabeling hospitalized patients represents a cost-saving measure for treating patients labeled as allergic to beta-lactams.

The importance of clinical history in the diagnosis of drug hypersensitivity in children

BACKGROUND

In case of suspected hypersensitivity reactions (HRs) to drugs, a challenging area for pediatricians is detecting relevant elements in the parent-reported history, in order to reach a definite diagnosis. We analyzed the concordance between the description of the HR and the medical reports documented at the time of the event. Furthermore, we studied any correlation between clinical history variables and the prediction of true allergy.

METHODS

We retrospectively collected 50 charts of children referred to our Allergy Unit, after a previous access to the Emergency Department. We compared the description of the HR at acute phase to the history told by parents. Type and timing of the HR and culprit drug were classified as "known" or "unknown." The diagnosis was confirmed or excluded at the end of the investigations. Logistic regression analysis was performed to find any significant association.

RESULTS

The type of the HR was known in 74%, the timing in 28%, and the culprit drug in 98%. We showed that having had a severe HR had an increased odds of remembering the timing; being older >6 years and having had an immediate HR had an increased odds of remembering the type; time to diagnostic was lower in patients whose parents remembered the type of HR.

CONCLUSION

Our paper underlines the importance of an accurate anamnesis at the time of the event. Providing the physicians with a standardized Case Report Form could be a useful tool to simplify the diagnostic work-up and minimize mistakes due to lack of memory.

IFN-γ ELISpot-enabled machine learning for culprit drug identification in nonimmediate drug hypersensitivity

BACKGROUND

Diagnosing drug-induced allergy, especially nonimmediate phenotypes, is challenging. Incorrect classifications have unwanted consequences.

OBJECTIVE

We sought to evaluate the diagnostic utility of IFN-γ ELISpot and clinical parameters in predicting drug-induced nonimmediate hypersensitivity using machine learning.

METHODS

The study recruited 393 patients. A positive patch test or drug provocation test (DPT) was used to define positive drug hypersensitivity. Various clinical factors were considered in developing random forest (RF) and logistic regression (LR) models. Performances were compared against the IFN-γ ELISpot-only model.

RESULTS

Among the 102 patients who had 164 DPTs, most patients had severe cutaneous adverse reactions (35/102, 34.3%) and maculopapular exanthems (33/102, 32.4%). Common suspected drugs were antituberculosis drugs (46/164, 28.1%) and β-lactams (42/164, 25.6%). Mean (SD) age of patients with DPT was 52.7 (20.8) years. IFN-γ ELISpot, fixed drug eruption, Naranjo categories, and nonsteroidal anti-inflammatory drugs were the most important features in all developed models. The RF and LR models had higher discriminating abilities. An IFN-γ ELISpot cutoff value of 16.0 spot-forming cells/106 PBMCs achieved 94.8% specificity and 57.1% sensitivity. Depending on clinical needs, optimal cutoff values for RF and LR models can be chosen to achieve either high specificity (0.41 for 96.1% specificity and 0.52 for 97.4% specificity, respectively) or high sensitivity (0.26 for 78.6% sensitivity and 0.37 for 71.4% sensitivity, respectively).

CONCLUSIONS

IFN-γ ELISpot assay was valuable in identifying culprit drugs, whether used individually or incorporated in a prediction model. Performances of RF and LR models were comparable. Additional test datasets with DPT would be helpful to validate the model further.

Penicillin Allergy Impact and Management

There is international evidence that penicillin allergies are associated with inferior prescribing and patient outcomes. A host of tools now exist from assessment (risk assessment tools, clinical decision rules) to delabeling (the removal of a beta-lactam allergy via testing or medical reconciliation) to reduce the impact of these "labels" in the hospital and community setting, as a primary antimicrobial stewardship intervention.

Approach for delabeling beta-lactam allergy in children

A considerable number of pediatric patients treated with beta-lactam (BL) antibiotics develop delayed onset of skin rashes during the course of treatment. Although the most frequent cause of these symptoms is infectious, many cases are labeled as allergic reactions to these drugs. BL allergy labels could have a negative impact, as they imply avoidance of this group of drugs and the use of second-line antibiotics, leading to a potential increase in adverse effects and the utilization of less effective therapies. This constitutes a major public health concern and economic burden, as the use of broad-spectrum antibiotics can result in multidrug-resistant organisms and prolonged hospital stays. Therefore, it is crucial to delabel patients during childhood to avoid false labeling in adult life. Although the label of BL allergy is among the most frequent causes of allergy referral, its management remains controversial, and new diagnostic perspectives are changing the paradigm of managing BL allergies in children. Traditionally, drug provocation testing (DPT) was exclusively performed in patients who had previously obtained negative results from skin tests (STs). However, the sensitivity of STs is low, and the role of in vitro testing in the pediatric population is not well defined. Recent studies have demonstrated the safety of direct DPT without prior ST or serum tests for pediatric patients who report a low-risk reaction to BLs, which is cost-effective. However, there is still a debate on the optimal allergic workup to be performed in children with a benign immediate reaction and the management of children with severe cutaneous adverse drug reactions. In this review, we will discuss the impact of the label of BL allergy and the role of the different tools currently available to efficiently address BL allergy delabeling in children.

Direct challenges are the gold standard for most antibiotic allergy evaluations

Antibiotic allergies are frequently encountered in clinical practice, and delabeling of these allergies has individual and public health benefits. This review focuses on the evidence supporting graded challenges without preceding skin testing in adult and pediatric patients to the major groups of antibiotics including penicillins, cephalosporins, sulfamethoxazole, fluoroquinolones, tetracyclines, macrolides, metronidazole, carbapenems, and aztreonam. The cost savings, time savings, and evidence for performing graded challenges outside of an allergy/immunology office are also reviewed for graded challenges to penicillins.

Genetic association of beta-lactams-induced hypersensitivity reactions: A systematic review of genome-wide evidence and meta-analysis of candidate genes

Importance

Beta-lactams (BLs) are the most prescribed antibiotics, being the most frequent cause of drug allergy. However, the association between BL allergy and genetic variations is still unclear.

Objective

This systematic review and meta-analysis aimed to summarize the genetic effects of BL-induced hypersensitivity using existing evidence.

Methods

We searched PubMed, Medline, Scopus, EMBASE, CINAHL, and Cochrane Library from inception to September 15, 2022 with no language restriction. Genetic association studies investigating genetic variant/polymorphism and risk of drug-induced hypersensitivity reactions among individuals receiving BL-antibiotics were included. We excluded studies of acute interstitial nephritis, drug-induced liver injury, serum sickness, and isolated drug fever. Data were comprehensively synthesized and quality of study were assessed using STrengthening the Reporting of Genetic Association Studies (STREGA). The record screening, extraction and quality assessment were performed by two reviewers and discussions were made to resolve discrepancies. The effects of each variant were pooled and evaluated by modified Venice criteria.

Results

A total of 9276 records were identified, and 31 studies were eligible for inclusion. Twenty-seven were candidate-gene association studies (5416 cases and 5939 controls), while the others were next-generation sequencing (NGS) or genome-wide association studies (GWASs) (119 838 cases and 1 487 111 controls). Forty-nine polymorphisms were identified and most of them located in allergic reaction pathways. Meta-analyses of 15 candidate variants in a mixture of both immediate and non-immediate reactions revealed weak genetic effects of rs1801275 (8 studies; n = 1,560; odd ratio 0.73; 95%CI: 0.57-0.93) and rs20541 (4 studies; n = 1,482; odd ratio 1.34; 95%CI: 1.07-1.68) in IL4R and IL13, respectively. Results from GWASs and NGS identified, and confirmed associations in HLA regions including HLA-DRA, HLA-B, HLA-DQA, HLA-DRB1, and HLA-DRB3.

Conclusion

Our study summarized genetic evidence influencing BL-induced hypersensitivity and estimated effects of potential variants. We postulated that the genomic studies provide better insights to the mechanism of reactions and suggest potential effects of HLA Class II variants. However, results were inconsistent and unable to generalize in different settings. Further high-throughput studies with a well-defined function, epigenetic interaction, incorporated with clinical factors, would be beneficial for risk identification in BL-induced hypersensitivity.

The Dutch Working Party on Antibiotic Policy (SWAB) guideline for the approach to suspected antibiotic allergy

OBJECTIVES

Prudent handling of reported antibiotic allergy is an important aspect of antibiotic stewardship. The Dutch Working Party on Antibiotic Policy (SWAB) constituted a multidisciplinary expert committee to provide evidence-based recommendations for bedside decision making in antibiotic therapy in patients that report an antibiotic allergy.

METHODS

The guideline committee generated 12 key questions, most of which were population, intervention, comparison and outcome (PICO) questions relevant for both children and adults with suspected antibiotic allergy. For each question a systematic literature search was performed and reviewed for the best available evidence according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system. Quality of evidence was graded from very low to high and recommendations were formulated in structured discussions as strong or weak.

RESULTS

Sixty recommendations were provided for suspected allergy to beta-lactam antibiotics (BLA) and non-beta-lactam antibiotics (NBLA). Due to the absence of randomized controlled trials in this field, the underlying evidence was predominantly graded as low or very low. Available data supports that a detailed allergy history should always be performed and critically appraised. When cross-allergy between BLA groups is not to be expected due to absence of molecular similarity of the side chains, the patient can be safely exposed to the alternative BLA. An exception to this rule are severe delayed type reactions, in which reexposure to a BLA should only be considered after consultation of a multidisciplinary team.

CONCLUSIONS

Accumulated scientific data now supports a more liberal approach that better balances benefits of treatment with first choice and usually smaller spectrum antibiotics with appropriate avoidance of antibiotics in case of a truly high risk of a (severe) allergic reaction. In the Netherlands, a formal guideline was developed that provides recommendations for the approach towards suspected allergy to BLA and frequently used NBLA, thereby strongly supporting antimicrobial stewardship.

The Safety of the Direct Drug Provocation Test in Beta-Lactam Hypersensitivity in Children: A Systematic Review and Meta-Analysis

BACKGROUND

Direct drug provocation test (DPT) without prior skin testing (ST) has been investigated in children suspected of being at risk for beta-lactam (BL) hypersensitivity reaction (HSR). However, no systematic review and meta-analysis has investigated the efficacy and safety of direct DPT for BL-HSR in children.

OBJECTIVE

To investigate the prevalence of BL-HSR by direct DPT and the safety of direct DPT in children.

METHODS

We searched MEDLINE, EMBASE, Web of Science, and CINAHL from their inception to July 23, 2022, for studies that performed direct DPT in children with suspected BL-HSR, or for studies that performed DPT in all cases with ST results, but they ignored the ST results. The true prevalence was defined as the proportion of children who experienced an HSR during direct DPT. Safety was determined according to the proportion of children who developed a dangerous reaction following DPT.

RESULTS

Twenty-eight studies with 8,334 direct challenges were included. Fifteen studies included patients who presented with either immediate or nonimmediate HSR, and the majority of the index reactions were nonsevere. Amoxicillin/amoxicillin-clavulanic acid was the most commonly used during the DPT. The pooled prevalence of confirmed BL-HSR was 5.23% (95% CI 4.17-6.39; I² = 72%). Immediate and nonimmediate HSR were reported in 0.8% (95% CI 0.43-1.25; I² = 55.1%) and 3.69% (95% CI 2.66-4.87; I² = 79.77%), respectively. Severe reactions were found in 3 cases with the frequency of 0.036% (95% CI 0.012-0.112; I² = 0%).

CONCLUSIONS

The prevalence of BL-HSR by direct DPT was 5.23%, and the frequency of severe reactions from direct DPT was very low (0.036%). Our findings support direct DPT as a safe and effective delabeling tool in children with suspected nonsevere BL-HSR.

Cofactors of drug hypersensitivity-A monocenter retrospective analysis

Background

Drug hypersensitivity reactions (DHRs) are major medical problems that influence the treatment of patients by both under- and overdiagnosis. Still, little is known about the role of predisposing or protecting cofactors of DHR.

Objective

This study aims to determine drug-specific cofactors in patients with DHR.

Methods

Retrospective file chart analysis of inpatients with suspected DHR in our department between 2015 and 2020 was performed. Descriptive statistics and multiple logistic regression were conducted for the estimation and statistical interference.

Results

DHRs were suspected in 393 patients with 678 culprit drugs. In 183 cases, drug hypersensitivities were confirmed, mostly against nonopioid analgesic drugs and antibiotics. Multiple logistic regression analysis identified a positive association of antibiotic hypersensitivity with obesity [odds ratio (OR) 5.75, average marginal effect (AME) +24.4%] and age and a negative association with arterial hypertension, female sex, elevated immunoglobulin E (IgE), and allergic rhinitis. Hypersensitivity to nonopioid analgesics was associated with atopic dermatitis (OR 10.28, AME +28.5%), elevated IgE, and arterial hypertension.

Conclusions

Drug-specific cofactors of DHR include obesity for antibiotics and atopic dermatitis for nonopioid analgesics, the knowledge of which may improve the risk calculation for drug provocation tests.

Beta-lactam hypersensitivity in children: Frequency and risk factors

AIMS

To determine the frequency of an authentic β-lactam (BL) hypersensitivity (HS) amongst a large number of children and to identify clinical risk factors that predict this hypersensitivity.

METHODS

All children with suspected BL allergy were evaluated by skin tests (ST) with the suspected BL. A 1-day oral provocation test (OPT) was performed in children with negative ST. We defined an authentic BL-HS case if the child exhibited a positive ST or a positive OPT. Risk factors associated with BL-HS were assessed using a univariate analysis. Covariates showing a P-value <.2 were included in the multivariate logistic regression analysis to determine independent predictors.

RESULTS

A total of 354 patients reporting 368 suspected BL reactions were included. The diagnosis of BL-HS was established in 24 children (6.7%). All these children had a positive ST. OPT was performed in 30 patients and was negative in all of them. In 110 children with a negative ST, BL was tolerated. In the risk factors analysis, 164 children were included. Older age (>5 years) at the reaction (odds ratio = 1.11; 95% confidence interval, 1.01-1.22; P = .02) and BL administered (odds ratio = 7.7; 95% confidence interval, 2.76-21.8; P < .001) were significantly associated with authentic BL-HS.

CONCLUSION

BL-HS should be evaluated with an appropriate allergy work-up before strict prohibition is made. In addition, age of patient and BL involved can be used as predictive factors of developing BL-HS in this population.

Improving drug allergy label accuracy by supervised safety- and protocol-driven evaluation

Introduction: Drug allergies are often self-reported but of unknown accuracy. We carried out a prospective study to examine the utility and safety of formal allergology evaluation, and to identify factors associated with accurate drug allergy labels. Method: All patients who underwent drug allergy evaluation in our clinic during the study period were recruited. Baseline demographics, characteristics of index hypersensitivity reaction and outcomes of evaluation were recorded. Results: A total of 331 patients from March 2019 to June 2021 completed drug allergy evaluation to index drugs of concern. There were 123 (37%) male patients, and the mean age was 49 years (standard deviation 17). There were 170 beta-lactam antibiotics, 53 peri-operative drugs, 43 others, 38 non-steroidal anti-inflammatory drugs, and 27 non-beta-lactam antibiotic evaluations. Index reaction occurred within 5 years in 165 (50%) patients, with latency of less than 4 hours in 125 (38%) patients. The most common index reactions were rash, angioedema and urticaria. There were 57 (17%) evaluations stratified as low risk, 222 (67%) moderate risk, and 52 (16%) high risk based on multidisciplinary consensus. Allergy label was found to be false (negative drug evaluation) in 248 (75%) patients, while 16/237 (7%) skin tests, 44/331 (13%) in-clinic graded challenge, and 23/134 (17%) home prolonged challenges were positive (true drug allergy). The most common evaluation reactions were rash and urticaria. No cases of anaphylaxis were elicited. Conclusion: Seventy-five percent of drug allergy labels are inaccurate. Risk-stratified, protocolised allergy evaluation is safe. Prolonged drug challenge increases the sensitivity of drug allergy evaluation and should therefore be performed when indicated. Keywords: Drug allergy, drug hypersensitivity, graded challenge, prolonged drug provocation, skin testing

Antibiotic Allergy De-Labeling: A Pathway against Antibiotic Resistance

Antibiotics are one of the most frequently prescribed drugs. Unfortunately, they also are the most common cause for self-reported drug allergy, limiting the use of effective therapies. However, evidence shows that more than 90% of patients labeled as allergic to antibiotics are not allergic. Importantly, the label of antibiotic allergy, whether real or not, constitutes a major public health problem as it directly impacts antimicrobial stewardship: it has been associated with broad-spectrum antibiotic use, often resulting in the emergence of bacterial resistance. Therefore, an accurate diagnosis is crucial for de-labeling patients who claim to be allergic but are not really allergic. This review presents allergy methods for achieving successful antibiotic allergy de-labeling. Patient clinical history is often inaccurately reported, thus not being able to de-label most patients. In vitro testing offers a complementary approach but it shows limitations. Immunoassay for quantifying specific IgE is the most used one, although it gives low sensitivity and is limited to few betalactams. Basophil activation test is not validated and not available in all centers. Therefore, true de-labeling still relies on in vivo tests including drug provocation and/or skin tests, which are not risk-exempt and require specialized healthcare professionals for results interpretation and patient management. Moreover, differences on the pattern of antibiotic consumption cause differences in the diagnostic approach among different countries. A multidisciplinary approach is recommended to reduce the risks associated with the reported penicillin allergy label.

Standards for practical intravenous rapid drug desensitization & delabeling: A WAO committee statement

Drug hypersensitivity reactions (DHRs) to intravenous drugs can be severe and might leave patients and doctors in a difficult position where an essential treatment or intervention has to be suspended. Even if virtually any intravenous medication can potentially trigger a life-threatening DHR, chemotherapeutics, biologics, and antibiotics are amongst the intravenous drugs most frequently involved in these reactions. Admittedly, suspending such treatments may negatively impact the survival outcomes or the quality of life of affected patients.

Delabeling pathways and rapid drug desensitization (RDD) can help reactive patients stay on first-choice therapies instead of turning to less efficacious, less cost-effective, or more toxic alternatives. However, these are high-complexity and high-risk techniques, which usually need expert teams and allergy-specific techniques (skin testing, in vitro testing, drug provocation testing) to ensure safety, an accurate diagnosis, and personalized management. Unfortunately, there are significant inequalities within and among countries in access to allergy departments with the necessary expertise and resources to offer these techniques and tackle these DHRs optimally.

The main objective of this consensus document is to create a great benefit for patients worldwide by aiding allergists to expand the scope of their practice and support them with evidence, data, and experience from leading groups from around the globe.

This statement of the Drug Hypersensitivity Committee of the World Allergy Organization (WAO) aims to be a comprehensive practical guide on the technical aspects of implementing acute-onset intravenous hypersensitivity delabeling and RDD for a wide range of drugs. Thus, the manuscript does not only focus on clinical pathways. Instead, it also provides guidance on topics usually left unaddressed, namely, internal validation, continuous quality improvement, creating a healthy multidisciplinary environment, and redesigning care (including a specific supplemental section on a real-life example of how to design a dedicated space that can combine basic and complex diagnostic and therapeutic techniques in allergy).

A Framework for Augmented Intelligence in Allergy and Immunology Practice and Research—A Work Group Report of the AAAAI Health Informatics, Technology, and Education Committee

Artificial and augmented intelligence (AI) and machine learning (ML) methods are expanding into the health care space. Big data are increasingly used in patient care applications, diagnostics, and treatment decisions in allergy and immunology. How these technologies will be evaluated, approved, and assessed for their impact is an important consideration for researchers and practitioners alike. With the potential of ML, deep learning, natural language processing, and other assistive methods to redefine health care usage, a scaffold for the impact of AI technology on research and patient care in allergy and immunology is needed. An American Academy of Asthma Allergy and Immunology Health Information Technology and Education subcommittee workgroup was convened to perform a scoping review of AI within health care as well as the specialty of allergy and immunology to address impacts on allergy and immunology practice and research as well as potential challenges including education, AI governance, ethical and equity considerations, and potential opportunities for the specialty. There are numerous potential clinical applications of AI in allergy and immunology that range from disease diagnosis to multidimensional data reduction in electronic health records or immunologic datasets. For appropriate application and interpretation of AI, specialists should be involved in the design, validation, and implementation of AI in allergy and immunology. Challenges include incorporation of data science and bioinformatics into training of future allergists-immunologists.

Outcome of a de-labelling algorithm compared with results of penicillin (β-lactam) allergy testing

Background

Penicillin allergy labels frequently impede guideline-directed treatment with a penicillin or other β-lactam antibiotics. Despite presumed allergy, targeted questioning may indicate a low probability of sensitization and permit reasonably safe administration of the antibiotic in question. In this study, we evaluated a standardized algorithm aiming to differentiate non-allergic patients from those with true allergic β-lactam hypersensitivity.

Methods

We retrospectively applied a de-labelling algorithm in 800 consecutive patients with suspected β-lactam hypersensitivity. All had undergone complete allergy work-up permitting to definitely exclude or diagnose β-lactam allergy between 2009 and 2019.

Results

In 595 (74.4%) out of 800 cases evaluated, β-lactam allergy could be excluded by negative challenge testing. IgE-mediated anaphylaxis was diagnosed in 70 (8.7%) patients, delayed-type hypersensitivity in 135 (16.9%). In 62 (88.6%) anaphylaxis cases, the algorithm correctly advised to use an alternative antibiotic. Accuracy was higher in patients with moderate to severe anaphylaxis (97.7%) compared to those with a history of mild reactions (73.1%). The algorithm correctly identified 122 (90.4%) patients with proven delayed-type hypersensitivity. It permitted de-labelling in 330 (55.5%) out of 595 patients with diagnostic exclusion of penicillin hypersensitivity, but failed to identify the remaining 265 (44.5%) as low-risk cases.

Conclusions

The algorithm detected 89.8% of cases with penicillin (β-lactam) allergy, sensitivity was optimal for moderate to severe anaphylaxis. Study data justify the implementation of a standardized de-labelling algorithm under close supervision in order to permit guideline-directed treatment and reduce the use of broad-spectrum antibiotics as part of an antibiotic stewardship program.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13223-022-00659-1.

New diagnostıc perspectives in the management of pediatrıc beta-lactam allergy

Since overdiagnosis of beta-lactam (BL) allergy is common in the pediatric population, delabeling is a critical part of antimicrobial stewardship. Undesirable consequences of inaccurate BL allergy labeling can be handled by incorporating traditional delabeling or newer risk-based strategies into antibiotic stewardship programs. Conventional assessment of BL allergy relies upon a stepwise algorithm including a clinical history with skin testing followed by drug provocation tests (DPTs). However, a growing number of studies highlighted the suboptimal diagnostic value of skin testing in children. Recently, there has been a paradigm shift in the practice of BL allergy assessment due to recent challenging data which emphasize the safety and accuracy of direct DPTs in children with a suspicion of non-immediate mild cutaneous reactions such as maculopapular eruption, delayed urticaria, and possibly also for benign immediate reactions such as urticaria/angioedema. Identifying low-risk BL allergy patients, in whom skin tests can be skipped and proceeding directly to DPTs could be safe, has become a hot topic in recent years. New risk stratification and predictive modeling studies that have the potential to better predict BL allergy risk status have recently been introduced into the field of drug allergy, particularly in adults. However, in contrast to adults, risk assessment studies in children are rare, and optimal risk definitions are controversial. In the coming years, promising potential methods to elucidate the predictors of BL allergy in children will require multidimensional approaches that may include predictive analytics, artificial intelligence techniques, and point-of-care clinical decision tools.

The effects of beta-lactam allergy relabeling on antibiotic prescribing practices

BACKGROUND

Beta-lactam antibiotic allergy labels are highly prevalent but rarely indicate an allergic intolerance. These patient-reported allergies lead to broad-spectrum antibiotic use, conferred resistance, increased expense, and adverse effects.

OBJECTIVE

To implement and assess the impact of a history-based clinical guideline that directs antibiotic management and beta-lactam allergy relabeling for patients reporting beta-lactam allergies.

METHODS

Patients with beta-lactam allergy labels were identified by our trained multidisciplinary team in diverse clinical settings. This quality improvement project was conducted to evaluate the safety and impact of the guideline on antibiotic use by comparing prescribing practices before and after guideline implementation.

RESULTS

A total of 79 patients with beta-lactam allergies were identified (penicillins-90%, cephalosporins-10%). After guideline implementation, outcomes of allergy relabeling included the following: (1) complete removal, indicating an unlikely true allergy (27%); (2) updated to detail successfully tolerated beta-lactam courses (48%); or (3) no change made to current label (25%). Beta-lactam antibiotic courses before and after guideline implementation compared with total antibiotic courses: allergy removed (44% vs 70%; P < .001), allergy updated (32% vs 68%; P < .001), and no change (27% vs 41%; P = .08). Compared with before guideline implementation, beta-lactam antibiotics were 3 times more likely to be prescribed after allergy assessment (odds ratio, 3.22; 95% confidence interval, 2.4-4.3; P < .05).

CONCLUSION

The implementation of the beta-lactam allergy clinical guideline resulted in most patients' allergy labels being removed or advantageously updated. These allergy label changes correlated with a substantial increase in the percentage of beta-lactam antibiotics prescribed. After guideline implementation, beta-lactam antibiotics had a 3-fold increased odds of being prescribed independent of allergy label outcome.

Key use cases for artificial intelligence to reduce the frequency of adverse drug events: a scoping review

Adverse drug events (ADEs) represent one of the most prevalent types of health-care-related harm, and there is substantial room for improvement in the way that they are currently predicted and detected. We conducted a scoping review to identify key use cases in which artificial intelligence (AI) could be leveraged to reduce the frequency of ADEs. We focused on modern machine learning techniques and natural language processing. 78 articles were included in the scoping review. Studies were heterogeneous and applied various AI techniques covering a wide range of medications and ADEs. We identified several key use cases in which AI could contribute to reducing the frequency and consequences of ADEs, through prediction to prevent ADEs and early detection to mitigate the effects. Most studies (73 [94%] of 78) assessed technical algorithm performance, and few studies evaluated the use of AI in clinical settings. Most articles (58 [74%] of 78) were published within the past 5 years, highlighting an emerging area of study. Availability of new types of data, such as genetic information, and access to unstructured clinical notes might further advance the field.

A combined risk modeling strategy for clinical prediction of beta-lactam allergies in children

Background: Drug provocation test (DPT) without skin tests is increasingly recommended in the evaluation of children with low-risk beta-lactam (BL) allergies. However, risk definitions are unclear. Objective: The aim of this study was to compose a clinical predictive model that could identify the children at low risk who could safely undergo direct DPT. Methods: The clinical data of 204 children who underwent a full diagnostic algorithm for suspected BL allergy were analyzed. Clinical data were used to construct mathematical predictive model for confirmed BL allergies. A prospective new sample was used for external validation of the final model. Results: The presentations during the index reaction were anaphylaxis in 5.9% and cutaneous reactions in the majority. BL allergy was confirmed in 15.7% of suspected cases. A backward multiple logistic regression model showed that a family history of drug allergy (adjusted odds ratio [aOR], 5.52), anaphylaxis (aOR, 5.14), any atopic disease other than asthma (aOR, 4.38), and a reaction interval of 0-6 hours during the index reaction (aOR, 5.32) were significantly associated with a confirmed BL allergy. A mathematical combined model based on these factors showed a sensitivity of 77.8% and a negative predictive value (NPV) of 94.3%. The validation study replicated sensitivity and NPV values of the main cohort. Conclusion: The risk definition in BL allergies should depend on population-specific predictive models, including a combination of significant risk factors rather than empiric risk approaches. This may help to accurately determinate children at low risk who may safely proceed to direct DPT.

Role of clinical history in beta-lactam hypersensitivity

PURPOSE OF REVIEW

Understand how the clinical history has been used to risk stratify patients reporting a beta-lactam allergy, both in clinical care pathways and predictive models.

RECENT FINDINGS

Drug allergy clinical care pathways have emerged as a safe and effective method of stratifying patients with a reported beta-lactam allergy into risk categories, with 'low-risk' patients able to proceed straight to direct challenges or test doses. These methods have streamlined antibiotic stewardship policies and penicillin allergy de-labeling. However, how to define 'low-risk' has been subject to much debate. New research has developed predictive models that utilize the clinical history to assess a patient's true risk of beta-lactam allergy.

SUMMARY

The clinical history has long been an essential part of drug allergy evaluation and has proven invaluable within the past decade in the development of drug allergy clinical pathways. Evidence-based predictive models that use the clinical history to assess a patient's true risk of beta-lactam allergy offer tremendous promise, but differ in crucial areas such as the populations they study, the predictor variables they use, and the ultimate accuracy they attain. These models highlight key aspects of the drug allergy history and pave the way for future large-scale research.

Predictive factors of amoxicillin immediate hypersensitivity and validation of PEN-FAST clinical decision rule

BACKGROUND

The challenge of delabeling amoxicillin allergy is an important issue for patients and clinicians, especially when anaphylaxis is reported. A recent study has proposed a clinical decision rule, PEN-FAST, to identify low-risk penicillin allergies.

OBJECTIVE

To validate the PEN-FAST clinical decision rule in a population with high risk of suspected immediate amoxicillin allergy and to identify clinical predictive factors of amoxicillin immediate hypersensitivity.

METHODS

We retrospectively analyzed medical records of patients with a suspected immediate amoxicillin allergy who carried out an allergologic evaluation by a specialist in the Allergy Unit of Strasbourg University Hospital from 2015 to 2020.

RESULTS

A total of 142 adult patients (88 women [62.0%]; median age, 52 [interquartile range, 40.3-62.0] years) were analyzed. Most of them reported anaphylaxis (68.8%). Internal validation of PEN-FAST score revealed a good discrimination with area under the curve of 0.86 (95% confidence interval, 0.79-0.92). A cutoff of less than 3 points for PEN-FAST was used to classify 29 from 142 patients at low risk of allergy, of whom only 2 (6.9%) received positive results of allergy testing. The negative predictive value for successful delabeling was 0.93 (95% confidence interval, 0.77-0.99). Predictive clinical features for immediate amoxicillin hypersensitivity were time since reaction (P < .001), time elapsed between drug intake and first symptom (P < .001), severity grade reaction (P < .001), and treatment or hospitalization required (P < .001).

CONCLUSION

PEN-FAST has been validated to identify low-risk penicillin allergies in our European cohort of patients mainly reporting anaphylaxis. This is the first reported external validation of a penicillin allergy clinical decision rule internationally.

Urticaria: The 1-1-1 Criterion for Optimized Risk Atratification in β-Lactam Allergy Delabeling

BACKGROUND

A spurious label of β-lactam allergy compromises antibiotic stewardship. Delabeling protocols based on direct challenges (ie, not preceded by allergy tests) can be applied in low-risk patients.

OBJECTIVE

This study aims at determining the significance of the characteristics of urticaria in the risk stratification for delabeling.

METHODS

The characteristics of urticarial eruptions that had occurred during therapeutic courses with a β-lactam, namely the time interval between the exposure and onset, the dose (first or subsequent) after which urticaria appeared, and the duration of the eruption, were correlated to the results of a systematic allergy workup (skin tests, specific IgE measurements, and challenges). Data from 410 patients enrolled in 3 allergy centers (Rome and Troina, Italy, and Antwerp, Belgium) were analyzed. A multivariable logistic regression was performed, which included appearance within 1 hour after the first dose and regression within 1 day: a model that can be summarized as the "1-1-1" urticaria criterion.

RESULTS

An urticarial eruption that had appeared within 1 hour after the first dose and had regressed within 1 day was more frequently reported in the group with a positive allergy workup, with odds ratios of 17 (95% confidence interval [CI]: 9-31), 11 (95% CI: 6-20), and 48 (95% CI: 14-157), respectively (P < .005). The 1-1-1 criterion displayed a sensitivity and specificity of 85%, and a negative predictive value and a positive predictive value of 80% and 90%, respectively.

CONCLUSION

Patients with urticaria meeting the 1-1-1 criterion should be considered at high risk and referred for an allergy workup with skin testing and specific IgE measurement before challenging.

Drug Allergy Profile From a National Drug Allergy Registry

Background: Drug hypersensitivity reactions (DHRs) are among the most frequent reasons for consultation in allergy departments and are becoming more common due to increasing prevalence and case complexity.

Objective: To describe the most common drugs associated with clinical reactions, diagnostic methods used, and outcomes of allergic evaluations of a national drug allergy registry over a 12-year period were used.

Methods: An observational, prospective, patient’s data registry-based study was conducted to analyze all referrals to the drug allergy outpatient clinics at Al-Rashed Allergy Center, Kuwait, between 2007 and 2019. Demographics, description of DHRs, and results of allergy tests to potential causative medications were reviewed. Diagnostic methods were focused mainly on skin tests (STs) and drug provocation test (DPT), when indicated.

Results: We evaluated 1,553 patients with reported DHRs. The mean age of the population was 41.52 ± 16.93 years, and the study population consisted of 63.7% female patients. Hypersensitivity was finally confirmed in 645 (41.5%) of patients, probable in 199 (12.8%), and not confirmed/nonallergic in 709 (45.6%) patients. Anti-inflammatory drugs and analgesics contributed to 39.22% of all confirmed drug allergies, followed by antibiotics 38.1% (β-lactam antibiotics (BLs) constituted 73.98% of all antibiotics and 28.21% of all drugs), anesthetics 1.8%, and radio-contrast media 0.31%. The majority of reactions were non-immediate 51.44%. The most commonly presenting symptoms among confirmed patients were urticaria 57.80%, angioedema 42.50%, respiratory symptoms 47.60%, and erythema 33.60%. Symptoms of anaphylaxis/anaphylactic shock were reported by 284 patients (44.00%) among confirmed cases. The most common method of diagnosis was a positive clinical history (54.4% in BLs and 90.45% in nonsteroidal anti-inflammatory drugs (NSAIDs). Among confirmed allergy to BLs, a positive ST was obtained in 31.9% of patients and positive DPT in 13.7%.

Conclusion: NSAIDs and antibiotics, mainly BLs, are the most commonly implicated in confirmed allergy. In both confirmed and not confirmed/nonallergic cases, BLs are the most frequently involved DHRs which are mainly immediate, and the most commonly presenting symptoms were urticaria, angioedema, and respiratory symptoms. Diagnosis was confirmed mainly by a positive clinical history and when indicated, by positive STs or a DPT.

Usefulness of an Artificial Neural Network in the Prediction of β-Lactam Allergy

BACKGROUND

An accurate diagnosis of β-lactam (BL) allergy improves the use of antibiotics, increases patients' safety, and reduces costs to health systems. Nevertheless, it requires skin and drug provocation tests, which are time-consuming and put the patient at risk. Furthermore, allergy testing is not available in circumstances such as the urgent need for antibiotic therapy.

OBJECTIVE

To evaluate the usefulness of an artificial neural network (ANN) in the prediction of hypersensitivity to BLs, and compare it with logistic regression (LR) analysis.

METHODS

In a single-center study, 656 patients evaluated for BL allergy between 1994 and 2000 were retrospectively analyzed, and the data were used to construct an ANN. The ANN predictive capabilities were compared with LR and then prospectively evaluated in 615 patients who underwent BL evaluation between 2011 and 2017.

RESULTS

A total of 1271 patients were evaluated. All patients had a definite diagnosis as allergic or nonallergic to BL. The prospective sample showed a lower percentage of patients with allergy than the retrospective sample (20.7% vs 25.8%; P = .018). In the retrospective and prospective series, the ANN reached a sensitivity of 89.5% and 81.1%, a specificity of 86.1% and 97.9%, a positive predictive value of 82.1% and 91.1%, and a negative predictive value of 92.1% and 95.2%, respectively. The ANN's performance was far superior to that of the LR, whose best performance reached a sensitivity of 31.9% and a specificity of 98.8%.

CONCLUSIONS

This ANN demonstrated a superior performance than the LR in predicting BL hypersensitivity without misdiagnosing severe allergic reactions. The ANN could be a helpful tool to classify the reaction risk, particularly in the identification of low-risk patients, in which an open challenge could be done to delabel patients.

Towards a more precise diagnosis of hypersensitivity to beta-lactams - an EAACI position paper

A recent survey of the European Academy of Allergy and Clinical Immunology (EAACI) Drug Allergy Interest Group (DAIG) on how European allergy specialists deal with beta-lactam (BL) hypersensitivity demonstrated a significant heterogeneity in current practice, suggesting the need to review and update existing EAACI guidelines in order to make the diagnostic procedures as safe and accurate, but also as cost-effective, as possible. For this purpose, a bibliographic search on large studies regarding BL hypersensitivity diagnosis was performed by an EAACI task force, which reviewed and evaluated the literature data using the GRADE system for quality of evidence and strength of recommendation. The updated guidelines provide a risk stratification in BL hypersensitivity according to index reaction(s), as well as an algorithmic approach, based on cross-reactivity studies, in patients with a suspicion of BL hypersensitivity and an immediate need for antibiotic therapy, when referral to an allergist is not feasible. Furthermore, the update addresses availability and concentrations of skin test (ST) reagents, ST and drug provocation test (DPT) protocols, and diagnostic algorithms and administration of alternative BL in allergic subjects. Specifically, distinct diagnostic algorithms are suggested depending on risk stratification of the patient into high and low risk based on the morphology and chronology of the reaction, immediate (ie, occurring within 1-6 hours after the last administered dose) or nonimmediate (ie, occurring more than 1 hour after the initial drug administration), and the reaction severity. Regarding the allergy workup, the main novelty of this document is the fact that in some low-risk nonimmediate reactions ST are not mandatory, especially in children. For DPT, further studies are necessary to provide data supporting the standardization of protocols, especially of those regarding nonimmediate reactions, for which there is currently no consensus.

Development and Validation of a Penicillin Allergy Clinical Decision Rule

IMPORTANCE

Penicillin allergy is a significant public health issue for patients, antimicrobial stewardship programs, and health services. Validated clinical decision rules are urgently needed to identify low-risk penicillin allergies that potentially do not require penicillin skin testing by a specialist.

OBJECTIVE

To develop and validate a penicillin allergy clinical decision rule that enables point-of-care risk assessment of patient-reported penicillin allergies.

DESIGN, SETTING, AND PARTICIPANTS

In this diagnostic study, a multicenter prospective antibiotic allergy-tested cohort of 622 patients from 2 tertiary care sites in Melbourne, Australia (Austin Health and Peter MacCallum Cancer Centre) was used for derivation and internal validation of a penicillin allergy decision rule. Backward stepwise logistic regression was used to derive the model, including clinical variables predictive of a positive penicillin allergy test result. Internal validation of the final model used bootstrapped samples and the model scoring derived from the coefficients. External validation was performed in retrospective penicillin allergy-tested cohorts consisting of 945 patients from Sydney and Perth, Australia, and Nashville, Tennessee. Patients who reported a penicillin allergy underwent penicillin allergy testing using skin prick, intradermal, or patch testing and/or oral challenge (direct or after skin testing). Data were collected from June 26, 2008, to June 3, 2019, and analyzed from January 9 to 12, 2019.

MAIN OUTCOMES AND MEASURES

The primary outcome for the model was any positive result of penicillin allergy testing performed during outpatient or inpatient assessment.

RESULTS

From an internal derivation and validation cohort of 622 patients (367 female [59.0%]; median age, 60 [interquartile range{IQR}, 48-71] years) and an external validation cohort of 945 patients (662 female [70.1%]; median age, 55 [IQR, 38-68] years), the 4 features associated with a positive penicillin allergy test result on multivariable analysis were summarized in the mnemonic PEN-FAST: penicillin allergy, five or fewer years ago, anaphylaxis/angioedema, severe cutaneous adverse reaction (SCAR), and treatment required for allergy episode. The major criteria included an allergy event occurring 5 or fewer years ago (2 points) and anaphylaxis/angioedema or SCAR (2 points); the minor criterion (1 point), treatment required for an allergy episode. Internal validation showed minimal mean optimism of 0.003 with internally validated area under the curve of 0.805. A cutoff of less than 3 points for PEN-FAST was chosen to classify a low risk of penicillin allergy, for which only 17 of 460 patients (3.7%) had positive results of allergy testing, with a negative predictive value of 96.3% (95% CI, 94.1%-97.8%). External validation resulted in similar findings.

CONCLUSIONS AND RELEVANCE

In this study, PEN-FAST was found to be a simple rule that accurately identified low-risk penicillin allergies that do not require formal allergy testing. The results suggest that a PEN-FAST score of less than 3, associated with a high negative predictive value, could be used by clinicians and antimicrobial stewardship programs to identify low-risk penicillin allergies at the point of care.

Recent developments and highlights in drug hypersensitivity

Drug hypersensitivity reactions (DHRs) are nowadays the third cause of allergy after rhinitis and asthma with a significant increase in prevalence in both adults and paediatric population with new drugs included as culprit. For this, DHRs represent not only a health problem but also a significant financial burden for affected individuals and health systems. Mislabelling DHRs is showing to be a relevant problem for both, false label of drug allergic and false label of nonallergic. All this reinforces the need to improve accurate diagnostic approaches that allow an appropriate management. Moreover, there is a need for training both, nonallergist stakeholders and patients to improve the reaction identification and therefore decrease the mislabelling. The use of allergy cards has shown to be relevant to avoid the induction of DHRs due to the prescription of wrong medication. Recent developments over the last 2 years and highlights about risk factors, diagnostic approaches, mechanisms involved as well as prevention actions, and management have been reviewed. In these papers, it has been outlined the need for correct diagnosis and de-labelling of patients previously false-reported as allergic, which will improve the management and treatment of patients with DHRs.

Short and extended provocation tests have similar negative predictive value in non-immediate hypersensitivity to beta-lactams in children

INTRODUCTION AND OBJECTIVES

Drug provocation tests (DPTs) are the gold-standard method to diagnose non-immediate hypersensitivity reactions (NIHSR) to beta-lactam antibiotics (BL) in children. Our aim was to compare the negative predictive value (NPV) of one-day (short) DPT versus 3-7 days (extended) DPT for the diagnosis of NIHSR to BL in paediatric age. A secondary aim was to compare confidence on drug re-exposure after short and extended negative DPTs.

METHODS

The occurrence of HSR on drug re-exposure and drug refusal after negative diagnostic DPTs were evaluated in children/adolescents with a history of NIHSR to BL using a questionnaire performed six months to ten years after DPT. Patients were divided into two groups according to the protocol performed: short DPT vs. extended DPT.

RESULTS

We enrolled 212 children and adolescents (86 females, 126 males, mean age at DPT 5.52 years, p25=3 years, p75=7.25 years): 69 tested with short DPT, and 143 with extended DPT. The NPV of both types of DPT together was 95.2%. The NPV of short DPT was 97.5% and the NPV of extended DPT was 93.8% (p=0.419). After negative DPT, beta-lactams were refused by carers in 14.75% of the children requiring subsequent treatment, 6.98% in the short DPT group and 18.99% in the extended DPT group (p=0.074).

CONCLUSIONS

In our paediatric sample, prolonging drug administration did not increase the NPV of diagnostic DPT for NIHSR to BL or reduce drug refusal. Altogether, the data here reported suggest that, however intuitive, prolonging DPT is not beneficial in the parameters analysed.

A systematic review shows no performance benefit of machine learning over logistic regression for clinical prediction models

OBJECTIVES

The objective of this study was to compare performance of logistic regression (LR) with machine learning (ML) for clinical prediction modeling in the literature.

STUDY DESIGN AND SETTING

We conducted a Medline literature search (1/2016 to 8/2017) and extracted comparisons between LR and ML models for binary outcomes.

RESULTS

We included 71 of 927 studies. The median sample size was 1,250 (range 72-3,994,872), with 19 predictors considered (range 5-563) and eight events per predictor (range 0.3-6,697). The most common ML methods were classification trees, random forests, artificial neural networks, and support vector machines. In 48 (68%) studies, we observed potential bias in the validation procedures. Sixty-four (90%) studies used the area under the receiver operating characteristic curve (AUC) to assess discrimination. Calibration was not addressed in 56 (79%) studies. We identified 282 comparisons between an LR and ML model (AUC range, 0.52-0.99). For 145 comparisons at low risk of bias, the difference in logit(AUC) between LR and ML was 0.00 (95% confidence interval, -0.18 to 0.18). For 137 comparisons at high risk of bias, logit(AUC) was 0.34 (0.20-0.47) higher for ML.

CONCLUSION

We found no evidence of superior performance of ML over LR. Improvements in methodology and reporting are needed for studies that compare modeling algorithms.

Outcomes from an inpatient beta-lactam allergy guideline across a large US health system

Objective:

To assess the safety of, and subsequent allergy documentation associated with, an antimicrobial stewardship intervention consisting of test-dose challenge procedures prompted by an electronic guideline for hospitalized patients with reported β-lactam allergies.

Design:

Retrospective cohort study.

Setting:

Large healthcare system consisting of 2 academic and 3 community acute-care hospitals between April 2016 and December 2017.

Methods:

We evaluated β-lactam antibiotic test-dose outcomes, including adverse drug reactions (ADRs), hypersensitivity reactions (HSRs), and electronic health record (EHR) allergy record updates. HSR predictors were examined using a multivariable logistic regression model. Modification of the EHR allergy record after test doses considered relevant allergy entries added, deleted, and/or specified.

Results:

We identified 1,046 test-doses: 809 (77%) to cephalosporins, 148 (14%) to penicillins, and 89 (9%) to carbapenems. Overall, 78 patients (7.5%; 95% confidence interval [CI], 5.9%–9.2%) had signs or symptoms of an ADR, and 40 (3.8%; 95% CI, 2.8%–5.2%) had confirmed HSRs. Most HSRs occurred at the second (ie, full-dose) step (68%) and required no treatment beyond drug discontinuation (58%); 3 HSR patients were treated with intramuscular epinephrine. Reported cephalosporin allergy history was associated with an increased odds of HSR (odds ratio [OR], 2.96; 95% CI, 1.34–6.58). Allergies were updated for 474 patients (45%), with records specified (82%), deleted (16%), and added (8%).

Conclusion:

This antimicrobial stewardship intervention using β-lactam test-dose procedures was safe. Overall, 3.8% of patients with β-lactam allergy histories had an HSR; cephalosporin allergy histories conferred a 3-fold increased risk. Encouraging EHR documentation might improve this safe, effective, and practical acute-care antibiotic stewardship tool.

Identifying Low-Risk Beta-Lactam Allergy Patients in a UK Tertiary Centre

BACKGROUND

There are marked geographical as well as temporal differences in patient sensitization profiles to β-lactams (BL).

OBJECTIVE

To determine the utility of skin test reagents and identify a cohort of patients where skin testing can be safely omitted in a cohort of patients referred to a UK tertiary referral center.

METHODS

A retrospective study of the clinical characteristics of 1092 patients referred for BL allergy testing was analyzed using multivariate regression analysis. The effectiveness of skin test reagents was also evaluated.

RESULTS

Multivariate logistic regression identified that a history of anaphylaxis (odds ratio [OR] 10.98, P = .001) and the patients' recall of the index drug (apart from ampicillin and meropenem, OR 3.51-12.43, P < .05) were independent predictors of type I BL allergic status and a time of less than 1 year elapsed since index reaction significantly increasing the odds of a patient with a history of anaphylaxis, having a type I BL allergy (OR 38.66, P = .003). An absence of anaphylactic severity, unknown name of the index drug and a reaction occurring more than 1 year before testing, has a negative predictive value (NPV) of 98.4%, which was similar to the NPV of skin testing of 98.9% for type I BL allergy. The NPV of skin testing with benzylpenicillin + amoxicillin ± index BL was similar with (98.9%) or without (98.1%) the use of benzylpenicillin polylysine and minor determinant for type I BL allergy.

CONCLUSION

We identified a "low risk" cohort of patients where the history is of similar reliability to skin testing in predicting nonallergic status for BL allergy.

Management of a surgical patient with a label of penicillin allergy: narrative review and consensus recommendations

Unsubstantiated penicillin-allergy labels are common in surgical patients, and can lead to significant harm through avoidance of best first-line prophylaxis of surgical site infections and increased infection with resistant bacterial strains. Up to 98% of penicillin-allergy labels are incorrect when tested. Because of the scarcity of trained allergists in all healthcare systems, only a minority of surgical patients have the opportunity to undergo testing and de-labelling before surgery. Testing pathways can be modified and shortened in selected patients. A variety of healthcare professionals can, with appropriate training and in collaboration with allergists, provide testing for selected patients. We review how patients might be assessed, the appropriate testing strategies that can be used, and the minimum standards of safe testing.