1.1 Introduction Acute cellulitis isa spreading infection of dermal and sub-dermal tissues with associatedinflammatory host response. It belongs to a group of conditions recently termedacute bacterial skin and skin structure infections (1), which are commonly encountered in EmergencyDepartments. Clinically, cellulitis manifests in a classic constellation ofdolor, calor, rubor and tumor over the skin. Involvement and inflammation ofneighbouring lymphatic tissues can progress to lymphagitis and regionallymphadenopathy (2).
Antibiotic therapy against gram-positivebacteria remains the cornerstone of treatment. In the course ofthis thesis, terms ‘cellulitis’ and ‘erysipelas’ will be joined to represent asingle clinical entity and referred to as cellulitis. This correlates with the USFDA terminology, which considers them synonymous due to the samepathopgysiological process behind these conditions (1). In the most recent IDSA nomenclatureupdate, both terms belong to a group of non-purulent SSTIs, which is used incurrent CPGs (2).
Practice of interchangeble use of theseterms is also common in European publications on the topic (3, 4). 1.2 Epidemology Cellulitis representsa significant burden to healthcare systems in the developed world and accountsfor between 1.5 and 3% of ED attendances in Europe and North America (5, 6). Admissions for thetreatment of cellulitis are second only to respiratory tract infection as themost common cause for inpatient antibiotic therapy in Europe (7). Approximately 12per 1,000 ED attendances in Ireland are due to cellulitis (8). In 2012, 12,184patients were admitted to Irish hospitals with cellulitis, of whom almost ahalf was admitted through the ED (9).
In the UnitedKingdom (UK), 87,749 people were admitted to hospital in 2010 for on average 7days with cellulitis, at a cost of up to £254 million to the exchequer (10). In the US, anestimated 14.5 million annual cases of cellulitis account for $3.7 billion inambulatory care alone (11). In addition, datafrom the US indicates that while the overall amount of annual visits for softtissue infections doubled in the recent years, the largest rate of increase invisits was experienced in the EDs, rising by 157% in 8 years (12). Although hospitaladmissions for ABSSSIs incur significant healthcare costs, they have been shownto represent only 7% of the overall share of cellulitis treated by EmergencyPhysicians and General Practitioners, with majority of patients beingdischarged home on oral treatment (4).
1.3 Management It is known that themost common microorganisms responsiblefor the majority of cellulitis cases belong to a group of ?-haemolyticstreptococci, while Staphylococcus Aureus is the second most common causativeagent (2, 11). Majority of current clinical guidelinesadvocate antibiotic treatment tailored against those pathogens, although thereis a significant variations in recommendations and practice (13, 14). The Clinical Resource Efficiency Support Team(CREST) guideline commonly used in UK and Ireland provides a cellulitisseverity classification system, with specific antibiotic type, dose and route ofadministration recommended for each class (6). By taking patient’sco-morbidities, early warning score and sepsis parameters into account, it alsoprovides a clear pathway to help clinician decide whether to admit a patientfor IV therapy or discharge home on a course of oral treatment table 1. Despite available guidelines and published research,there is no evidence-based consensus on a preferred antibiotic treatment forcellulitis at present. A recent Cochrane Review that examined 25 RandomizedControlled Trials (RCT) of cellulitis therapy, found no clearly superior,single treatment (15).
There is an obviousclinical equipoise between the use of oral flucloxacillin alone or combinedwith phenoxymethylpenicillin for the ED treatment of cellulitis as evidenced bycurrent disparate prescribing practice and hospital guidelines, with no RCT comparingthe two treatments conducted to date. Table 1: CRESTrecommendations. Guideline CREST Modified score Antibiotic therapy Class I No signs of systemic toxicity, have no uncontrolled co-morbidities and can usually be managed with oral antimicrobials on an outpatient basis. No comorbidity No sepsis SEWS < 4 Flucloxacillin 500mg qds po Penicillin allergy: Clarithromycin 500mg bd po Class II Either systemically ill or well but with a co-morbidity such as peripheral vascular disease, chronic venous insufficiency or morbid obesity. >1 comorbidity No sepsis SEWS < 4 Flucloxacillin 2g qds IV or Ceftriaxone 1g od IV (OPAT only) Penicillin allergy: Clarithromycin 500mg bd IV or Clindamycin 600mg tds IV Class III Significant systemic upset (acute confusion, tachycardia, tachypnoea, hypotension) or may have unstable co-morbidities that may interfere with a response to therapy or have a limb threatening infection due to vascular compromise.
Sepsis SEWS < 4 Flucloxacillin 2g qds IV Penicillin allergy: Clarithromycin 500mg bd IV or Clindamycin 900mg tds IV Class IV Sepsis syndrome or severe life threatening infection such as necrotizing fasciitis. Sepsis SEWS ? 4 Benzylpenicillin 2.4g 2-4 hourly IV + Ciprofloxacin 400mg bd IV + Clindamycin 900mg tds IV (If allergic to penicillin use Ciprofloxacin and Clindamycin only) 1.4 Incidence ofpenicillin allergy Penicillin, either as flucloxacillin and/or phenoxymethylpenicillin, isthe standard antibiotic regimen for the treatment of cellulitis in Ireland andthe UK and is also recommended by French and Norwegian prescribing guidelines (16, 17). With asignifficant proportion of cellulitis being seen in the ED as mentionedearlier, it becomes increasingly important for an EP to be informed about thepotential adverse effects of the reccomended treatment. Penicillin allergycan be defined as a reaction following an ingestion of a penicillin-basedantibiotic, with features suggestive of immunological mechanism (18).
It represents the largest share of allergicreactions attributable to antibiotics, while antibiotics constitute themajority of all drug allergy reports in the ED setting (19, 20). Despite the significant amount of informationavailable on the general epidemiology of PA, its incidence among ED patientswith acute cellulitis treated with penicillin antibiotics is understudied,leaving a knowledge gap that will be targeted by the first systematic review inthis thesis. We will also directly assess the incidence ofPA in the RCT at the basis of this thesis.
1.5 Lymphoedema and cellulitis. Lymphoedemais a chronic condition characterized by fluid retention and tissue swelling inthe affected extremity, which is caused by the compromise in lymphaticdrainage. A recent systematic review identified it is a unique risk factor forlower limb cellulitis (21), while smaller studies haveconcluded that impairment in lymphatic system is a common finding in patientswith acute cellulitis and erysipelas, as demonstrated by lymphoscintigraphy ofthe limb following resolution of the infection (22, 23). It is important to note that whilethe initial presence of lymphoedema predisposes to cellulitis by creating alocal immune deficiency, it is also known that cellulitis can result in damageto lymphatic system, subsequently leading to the development of secondarylymphoedema (24). Despite theabove-mentioned research, lymphoedema of the lower limb remains a relativelyunderstudied condition (25), and information on its prevalencein ED patients with acute cellulitis is currently lacking.
Data collected onpatients in our RCT will be analyzed in order to determine the prevalence oflymphoedema in this group of patients and assess the association betweenlymphoedema and the recurrence of cellulitis. 1.6 Adherence toantibiotic treatment Adherence tomedication is the degree to which a person’s drug-taking behavior correspondswith the agreed recommendations from a health care provider. It is acceptedthat poor patient adherence to medication is associated with higher rate oftreatment failure, more frequent return visits to a physician, prolongedhospitalization time and increased healthcare costs (26-28). In the context ofclinical trials, variable adherence to prescribed medication is an importantand recognised cause of fluctuation in drug response, which subsequentlyresults in unnecessary dose escalations, underestimation of efficiacy of thecompound, underestimated incidence of dose-dependent adverse effects andemergence of antibiotic-resistant pathogens (29).
Ultimately, this may lead tomisinterpretation of the trial results, type 2 errors in judging efficiacy anddistorted pharmacoeconomic analysis (30). At the same time, optimal adherence to bothmedication and placebo regimens is known to predict better patient outcomes,making documentation of adherence an essential component of modern clinicaltrials (28, 31). There are multiplemethods of measuring medication adherence used today, which are fundamentallydivided into direct and indirect, with each group having its advantages anddisadvantages.
Direct measurements of adherence involve a physician observing the patient during ingestion ofmedication, or use specific biochemical markers to estimate concentration of apharmaceutical agent in body fluids. Indirect measurements include, but are notlimmited to asking patients about their compliance either verbally or through awritten questionaire, counting pills, using pharmacy databases to determine therate of collecting prescriptions, utilizing electronic medication adherencemonitoring systems or assessing clinical response to treatment. Table 2summarises some of the commonly used methods. An increasing bodyof systematic evidence is pointing towards a superiority of electronicmedication adherence monitoring over other indirect methods in terms ofaccurancy, reliability and susceptibility to bias (26, 29, 32,33). While several techniques of electronicadherence monitoring exist, electronic detection of package entry is thecurrent gold standard of collecting adherence data in clinical trials (30, 33). The RCT at the basisof this thesis is going to utilize three indirect methods – patient self-reporting,pill count and electronic monitoring (medication event monitoring system MEMS®) in order tomeasure adherence to prescribed antibiotic therapy in ED patients with acutecellulitis. The correlation between these measures will then be assessed, andused to inform interpretation of the RCT’s clinical outcomes. Collected datawill be analyzed in order to determine the potential financial implications ofnon-adherence to penicillin antibiotics among ED patients with acute cellulitisin this trial.
We will also conduct a systematic review ofliterature, in order to determine how is antibiotic adherence measured in otherED-based clinical trials on cellulitis to date. Table 2: Common methods of monitoring adherence. Method Advantages Disadvantages Direct Directly observed therapy.
Most accurate Can be distorted by patients; impractical. Measuring level of medication or it’s metabolite in blood. Objective Affected by variations in metabolism and ‘white coat adherence’, expensive. Measuring biological marker in blood or body fluids. Objectivel; can be used to measure placebo Requires costly quantitative assays and collection of body fluids.
Indirect Patient questionaires, verbal reports or pill counts. Simple, cheap, most useful in clinical setting Easily distorted by patients, susceptible to error with prolonged time between visits. Pharmacy databases to verify prescription collections. Objective, easily accessible data. Does not quate to actual consumption of medication, requires a closed database. Electronic medication adherence monitoring systems. Quantifiable and precise, detects patters of medication consumption Requires upload of data from medication containers, expensive. Clinical response or chosen physiological parameters.
Simple, easy to conduct. Alternative factors may affect clinical response or interfere with physiology.