Protocol. This trial protocol has been provided by the authors to give readers additional information about their work.

Transcription

1 Protocol This trial protocol has been provided by the authors to give readers additional information about their work. Protocol for: Berende A, ter Hofstede HJM, Vos FJ, et al. Randomized trial of longer-term therapy for symptoms attributed to Lyme disease. N Engl J Med 2016;374: DOI: /NEJMoa

2 NEJM PLEASE A randomized, placebo- controlled trial comparing short- and long- term antibiotic treatment for Lyme borreliosis- attributed persistent symptoms Supplement for review purposes Protocol amendment overview This supplement contains the following items: 1. Original study protocol, Version 4.1, effective 24 MAR Letter to IRB Re. Protocol amendment 1, Protocol Version 4.2, effective 15 JUL Letter to IRB Re. Protocol amendment 2, Protocol Version 4.3, effective 01 MAR Overview of Protocol amendments 5. Final Protocol, Version 4.3, effective 01 MAR Statistical Analysis plan, first and final version 1.7, effective 7 MAR 2014 No SAP amendments have been made Study time lines Study protocol v.4.2 released : 15 JUL 2010 First patient enrolled : 03 SEP 2010 Last patient enrolled : 17 MAY 2013 Statistical analysis plan approved : 07 MAR 2014 Start, blinded data QC : 14 MAR 2014 Last patient f/u study visit : 01 JUN 2014 Blinded database locked : 13 OCT 2014 Data unblind : 14 OCT 2014 PLEASE Study Protocol Amendment overview April 29, 2015 Page 1

3 Original study protocol Version 4.1 Effective 24 MAR 2010 PLEASE Study Protocol Amendment overview April 29, 2015 Page 2

4 PLEASE Persistent Lyme Empiric Antibiotic Study Europe A prospective, randomised study comparing two prolonged oral antibiotic strategies after initial intravenous ceftriaxone therapy for patients with symptoms of proven or possible persistent Lyme disease Protocol version 4.1 ( ) Principal Investigator Prof BJ Kullberg, M D Division of infectious diseases Department of Medicine University Medical Center Nijmegen St Radboud P.O. Box HB Nijmegen The Netherlands Phone Fax B.kullberg@aig.umcn.nl Investigators A Berende, M.D. Infectious Diseases, UMCN Infectious diseases H ter Hofstede, M.D. Infectious Diseases, UMCN Infectious diseases F Vos, M.D. Infectious Diseases, UMCN, and St Infectious diseases Maartenskliniek, Nijmegen DSC Telgt, M.D. Infectious Diseases, UMCN, and St Infectious diseases Maartenskliniek, Nijmegen F van der Hoogen, M.D. Rheumatology, UMCN, and St Rheumatology Maartenskliniek, Nijmegen AWM Evers, Ph.D. Medical Psychology and Rheumatology, Psychology UMCN Prof J Galama, M.D. Medical Microbiology, UMCN Medical microbiology Prof MG Netea, M.D. Infectious Diseases, UMCN Infectious diseases LAB Joosten, Ph.D. Infectious Diseases, UMCN Infectious diseases Lyme protocol version MAR 2010 Page 1

5 Content 1. Summary Background Lyme Disease Diagnostic issues Treatment Study Rationale Persistent and post Lyme disease studies Diagnostic criteria Choice of antibiotics Assessments of outcome Study objective Primary objective Secondary objective Study design Informed consent & institutional review Criteria for subject selection Inclusion criteria Exclusion criteria Randomisation procedure Methodology Pre-randomisation (baseline) evaluation Neuropsychological and functional evaluation Study visits Accelerometers Follow up Ethical and regulatory considerations Study treatments Ceftriaxone Randomized study drug preparation Study drug supply and accounting Unblinding of randomization Concomitant medication Definition of efficacy endpoints Primary endpoint Secondary endpoints Safety Adverse events Serious adverse events Study drug intolerance and discontinuation from therapy Discontinuation from study Unblinding of randomisation Recording of data Statistical methodology Modified intent-to-treat analysis group Evaluable patients subgroup Statistics and sample size calculation Lyme protocol version MAR 2010 Page 2

6 15.4 Interim analyses Economic evaluation Study committees Independent safety committee Independent physician Publication of study results References Lyme protocol version MAR 2010 Page 3

7 1. Summary While diagnosis and treatment of acute Lyme disease is relatively easy, this is far more problematic for chronic or persistent Lyme disease (PLD). Many patients present with a history and complaints compatible with persistent Lyme disease. There is considerate controversy about the diagnosis of these patients, and diagnostic tests currently available cannot exclude or confirm the diagnosis active Lyme disease. These patients are increasingly demanding empirical treatment for possible or presumed persistent Lyme disease. As yet, there are no prospective randomized studies that have addressed prolonged therapy in either confirmed or possible persistent Lyme disease. Patients with possible PLD often present with musculoskeletal symptoms, with or without arthritis, paraesthesia, cognitive impairment and fatigue. A majority will receive a short course of antibiotics from their general practitioner but in the end, they will be referred to internal medicine specialist/infectious disease specialist, rheumatologists and sometimes neurologists for (further) antibiotic treatment. However, current serological tests are appropriate for early manifestations of Lyme disease, but hardly of any value for diagnosing PLD. This hampers decision-making about antibiotic treatment, which in turn leads to much controversy and high medical consumption. On the one hand, the diagnosis is rejected (and treatment withheld) in patients in whom PLD can eventually be proven. On the other hand, in many other patients, who do receive long-term antibiotic treatment, the diagnosis PLD seems very unlikely. Patients suffer from uncertainty concerning the diagnosis of PLD, and anxiety about possible disabling sequelae. At present, the scientific basis for long term antibiotic treatment for musculoskeletal symptoms related to possible PLD is lacking. This strategic study focuses on patients with chronic musculoskeletal symptoms of unknown origin, and possible PLD (possible or certain past infection with B. burgdorferi). All patients will be treated with ceftriaxone for 2 weeks and will then be randomised between three treatment strategies: (A) oral doxycycline (12 weeks); (B) clarithromycin and hydroxychloroquine (12 weeks); or (C) placebo (12 weeks). Patients in the placebo group will be offered antibiotic treatment after the study. The effect of treatment will be measured using strictly standardized medical and psychological examination. The primary endpoint will be the health-related quality of life after end of antibiotic treatment (week 14). In addition, neuropsychological evaluation and registration of physical activity with a movement-sensor will be done at three different time points. Prognostic factors (e.g. previous treatment, presence of specific antibodies and/or B. burgdorferi DNA) for outcome of antibiotic treatment will be measured as well. Aim of the study is to establish whether prolonged antibiotic treatment of patients diagnosed with presumed PLD (as endorsed by the international ILADS guidelines) leads to better patient outcome than short-term treatment as endorsed by the Dutch CBO guidelines. One can expect that this will help to determine which patients with longstanding disabling symptoms and suspected PLD will, and which patients will not benefit from chronic antibiotic treatment. Lyme protocol version MAR 2010 Page 4

8 2. Background As in other European countries and the US, Lyme disease is a growing problem in the Netherlands. In 1994, erythema migrans (EM), the typical skin rash which accounts for approximately 80% of cases of Lyme disease, was diagnosed in 6000 patients in the Netherlands. In 2005 this number had tripled to 17000, while in the same period, the number of medical consultations for tick bites increased from to A recent report has shown that the proportion of ticks infected with Borrelia has increased significantly, now averaging 24%, and up to 50% in several areas, further increasing the health burden. 2 In parallel with the growing incidence of acute Lyme disease, there is also a rapid increase in patients with chronic symptoms attributed to infection with B. burgdorferi. This (true or presumed) late or persistent Lyme disease (PLD) may follow EM or other (unnoticed) manifestations of acute Lyme disease, and may also appear in patients who have received adequate antibiotic treatment. Patients with PLD mainly present with pain, arthritis, musculoskeletal symptoms, fatigue, neurological and cognitive disturbances. 3;4 While acute Lyme disease is primarily managed by general practitioners, patients with presumed PLD are mostly referred to infectious disease specialists, neurologists, and rheumatologists. At present, an estimated patients are referred to Dutch hospitals yearly for PLD, and this number is still growing. The presentation of presumed PLD is variable, but the majority of patients with PLD suffer from longstanding and disabling symptoms. Pain of muscles and joints, arthritis, signs of encephalopathy and fatigue influences the daily life of these patients. 5 Chronic pain is an important contributor to impairment of physical health and was shown to be similar to that reported by patients with osteoarthritis. 5 Because PLD is difficult to diagnose with the present tools, there is much controversy about this disease, leading to a high consumption of medical resources. Currently available diagnostic tools (ELISA serology or Western Blot) are adequate for the diagnosis of early Lyme disease, but have little value for the diagnosis of PLD Nevertheless, serological tests are ordered repeatedly in local and even international laboratories, although their predictive value for indication or outcome of treatment outcome are unknown. Doctors experience increasing pressure of patient groups, demanding empirical treatment where current techniques fail to confirm the diagnosis of persistent infection. Many patients feel disowned, having persistent musculoskeletal symptoms, whereas a diagnosis of Lyme disease is rejected by their physicians by a putatively false-negative serological test. Indeed, we and others have demonstrated active persistent Lyme disease in a number of such patients, using experimental techniques (culture, immunofluorescence, PCR for Borrelia DNA in skin or synovial biopsies), in spite of prolonged previous antibiotic treatment. On the other hand, it is assumed that many patients referred with chronic symptoms and treated with long term antibiotics do not have PLD. Currently, the decision whether or not to treat is made arbitrarily, as evidence-based strategies are lacking. Therefore, the current approach is highly inefficient and costly. In addition, PLD has important socio-economic Lyme protocol version MAR 2010 Page 5

9 consequences. Patients are forced to interrupt, adapt or even give up their education or work, and direct costs of their health-seeking behaviour are high. There are only few prospective, controlled data to support chronic antibiotic treatment for PLD. Therefore, evidence for prescribing or withholding prolonged antibiotic treatment for PLD would end current controversy and lead to uniform and cost-effective management of this disease entity. If no beneficial effect can be shown, unnecessary antibiotic treatment can be withheld, and other therapeutic strategies (e.g. cognitive behavioural therapy) can be explored. 2.1 Lyme Disease Lyme disease is a tick-borne spirochete infection with variable signs and symptoms. Without treatment neurologic complaints, arthritis, atrophic skin disease or heart conduction problems may develop. In late stages of disease chronic encephalopathy may become manifest. In the United States, the disease is caused by the spirochete Borrelia burgdorferi sensu strictu. In Europe, other Borrelia spp. have been found: Borrelia afzelii and Borrelia garinii, causing differences in both clinical disease and diagnostic techniques required. In the Netherlands 25% of the ticks are infected with Borrelia, and the incidence of acute Lyme disease (EM) is estimated 4.3 per inhabitants, leading to approximately 9000 new patients with Lyme disease yearly. 18;19 Acute Lyme disease Although the characteristic skin rash, erythema chronicum migrans (EM) is considered pathognomonic for Lyme disease, only 75-90% of patients present with EM. 7 Consequently, the disease remains untreated in a considerable proportion of patients. Subsequently, patients may develop intermittent unilateral arthritis or neurologic signs, such as meningitis, neuritis, or radiculoneuritis. Neuroborreliosis may occur in 10-20% of non-treated patients. 15 Since Borrelia garinii is the most important species causing neurologic symptoms, neurologic disease is seen more often in Europe. 20;21 Persistent Lyme disease Chronic or late disease includes chronic joint symptoms, acrodermatitis chronica atrophicans (ACA), which is caused by Borrelia afzelii and is frequently diagnosed in Europe 21, as well as neurological signs and symptoms. Signs of encephalopathy (fatigue, cognitive dysfunction, memory impairment) and radiculoneuropathy or polyneuropathy may occur months to years after the acute disease. 6;12;17;22-26 There is considerate controversy about the diagnosis of the latter patients. Many authors have argued that these patients have other diseases, such as fibromyalgia or chronic fatigue syndrome. 8-10;27 Various possible mechanisms leading to a chronic syndrome of persistent Lyme disease have been postulated: (1) permanent tissue damage by the spirochete itself, (2) an immunological process caused by live Borrelia, or (3) persisting antigens of the dead Lyme protocol version MAR 2010 Page 6

10 microorganism leading to immune activation through molecular mimicry. 11;28;29 Positive polymerase chain reaction (PCR) for Borrelia DNA have been found in skin or synovial biopsies from patients with a clinical diagnosis of persistent Lyme disease not responding to multiple courses of antibiotics, strongly arguing for persistence of live B. burgdorferi infection causing chronic disease (UMCN, data on file). 2.2 Diagnostic issues Although isolating the microorganism is the gold standard of infectious disease, culture of Borrelia is cumbersome, especially from other sites than the skin. 7;14;30-32 PCR to detect DNA of Borrelia outer surface proteins (Osp) A, B and C and flagellin has been developed and validated in animal experiments as well as in skin biopsies of infected patients. 33;34 PCR has also been described on other body materials, such as synovial fluid, cerebrospinal fluid, and urine The specific immune response to Lyme disease develops gradually over a period of months. Differences in antigenic variation and expression between European and American strains and cross reactions may influence specificity and sensitivity of antibody assays. 20;21 Antibodies are positive in only 20-30% during EM and in 70% after 1-4 months. 7;13 Besides, antibody titers do not correlate with the stage of disease. 16 Persisting Lyme disease may be found in patients with negative serology after early treatment. 13;41 A Western blot may be used to confirm positive serology, although this inevitably leads to a further reduction of sensitiviy. 7;14;16 Moreover, no consensus has been reached on the optimal antigens to be included in the Western blot for sensu latu strains in Europe. 42 Measuring intrathecal antibody production is also potentially misleading due to low sensitivity. 6;14; Treatment Although spontaneous cure may occur, treatment is recommended for B. burgdorferi infection as the first series of 55 patients with EM in Lyme (US) developed arthritis in 50% and chronic synovitis in 10% when untreated. 46 Borrelia spp are sensitive to tetracycline, most macrolides, and to beta-lactam antibiotics such as penicillin and third generation cephalosporins. Most clinical studies have used amoxicillin, ceftriaxone, or doxycycline. For macrolides, results in animal models have been variable. 13;47-52 Erytromycin and doxycycline may be more effective against intracellular Borrelia than penicillin or ceftriaxone. 53 Borrelia spp. are insensitive to quinolones, first generation cephalosporines, trimethoprimsulfamethoxazole, rifampicin and aminoglycosides. 15;54 There is a considerable relapse rate (10-20%) after treatment with the currently recommended regimens in both acute and late stage Lyme disease, as well as in animal models of Lyme borreliosis. 13;55-58 Since Borrelia spp. show a slow intracellular growth and survive in less vascularised areas such as ligaments 59;60, prolonged exposure time rather Lyme protocol version MAR 2010 Page 7

11 than concentration of antibiotics correlates with elimination of Borrelia. 61 Especially in chronic symptoms or persistent Lyme disease, prolonged antibiotic treatment has been advocated. 17;39;49;52;55;58;62 Two different approaches to PLD are being advocated in the Netherlands: (1) The national CBO consensus for Lyme Disease (2004) does recommend treatment for a maximum of 2 weeks for most manifestations of Lyme disease: 2 weeks of ceftriaxone for established neuroborreliosis, and 2 weeks of doxycycline for most other types of Lyme disease 63. (2) In contrast, the international 'Evidence-based guidelines for management of Lyme borreliosis' by ILADS (International Lyme and Associated Diseases Society) 64, mainly supported and advocated by the Netherlands Society for Lyme Patients (NVLP) and used by a number of healthcare providers in the Netherlands, recommend long-term treatment for PLD, using ceftriaxone, doxycycline, macrolides or combination therapy for 3-6 months or more 64. In general, prolonged antibiotic therapy is not supported by physicians in the Netherlands, and many patients travel to specialised laboratories (e.g., in Cologne) and clinics (e.g., in New York and Switzerland) to seek medical help. It is clear that the burden of disease for these patients is high, and the co-existence of different and opposing treatment guidelines is confusing and worrisome, leading to a high consumption of medical resources. 3. Study Rationale 3.1 Persistent and post Lyme disease studies Lyme disease is a growing problem in the Netherlands. There is a growing incidence in of acute Lyme disease and also in patients with chronic symptoms attributed to infection (PLD). An estimated patients are referred to Dutch Hospitals yearly for true or presumed PLD. No consensus on approach or treatment exists for these patients, and presently, two opposing and conflicting guidelines are being used in clinical practice. Evidence for prescribing or withholding prolonged treatment for PLD would end the current controversy and lead to uniform and cost-effective management. If no beneficial effect can be shown, unnecessary antibiotic treatment can be withheld, and other therapeutic strategies (e.g. cognitive behavioral therapy) can be explored. With increased pressure from health authorities (VWS) and patients groups, CBO has reinstalled a Lyme guideline committee in 2007, specifically to re-evaluate the evidence on PLD. However, the committee has identified the lack of any prospective data on this subject, Lyme protocol version MAR 2010 Page 8

12 and the revised guideline will not be provided a definitive answer. This underscores that a study on the effectiveness of the two guidelines is urgently needed. There are few prospective, controlled data to support chronic antibiotic treatment for PLD. Several open studies have suggested favourable results of prolonged antimicrobial therapy for patients with PLD. The efficacy of ceftriaxone, 2 g per day for 30 days, was evaluated in 18 patients with presumed PLD who met strict criteria for Lyme encephalopathy. 65 After treatment, neuropsychological evaluation, CSF (central nervous system fluid) analyses and SPECT imaging were significantly improved in most patients. Also, all 18 patients rated themselves as back to normal or improved. 65 In a small open study, patients with persistent symptoms and either positive ELISA or a history of erythema migrans received combined therapy with roxitromycin and cotrimoxazole for 5 weeks, leading to complete resolution of symptoms in 76%. 49;50 Whereas prospective trials are lacking, multiple case series have suggested a beneficial effect of oral doxycycline in patients with presumed PLD. In an open study on 254 patients with Lyme-compatible symptoms, 90% were partially or completely cured after oral antibiotic treatment with doxycycline for 8-12 weeks 17, although we and others did not report such beneficial results. Most of these patients had a positive Western Blot against Borrelia antigens (80%) at inclusion, although their ELISA was negative in 52%. 17 In another open, retrospective study on 235 patients with suspected PLD, combination therapy of clarithromycin plus hydroxychloroquine resulted in 45% of patients being markedly improved after 3 months of treatment 3. However, none of these studies included a control arm. A large, prospective US study by Klempner et al compared ceftriaxone followed by doxycycline versus placebo for a total of 3 months, showing no beneficial effect, but this study has been subject of major criticism for several reasons. 5 First, the patient group included may have been inhomogeneous. Second, the Klempner study was terminated prematurely after interim analysis had indicated that it was unlikely that a significant difference in efficacy between treatment groups would be reached. Third, Lyme disease is caused by different Borrelia species in the US and Europe, with quite different clinical manifestations (e.g., monoarthritis in the US, versus polyarthralgias, myalgia, and neurological involvement in Europe), and US treatment studies cannot be generalised. Ours is the first prospective trial on PLD in Europe. Based on the above, it seems evident that PLD is an important and ever growing problem with significant physical, psychological and socio-economic consequences. Patients presenting with PLD are often prescribed long-term antibiotic treatment according to conflicting guidelines, but the scientific basis for this policy is weak, and the effectiveness of the opposing guidelines has not been compared. Lyme protocol version MAR 2010 Page 9

13 3.2 Diagnostic criteria Many patients present with a history and complaints compatible with persistent Lyme disease. There is considerable controversy about the diagnosis of these patients. 8-10;27;66 As mentioned in section 2.2, diagnostic tests currently available cannot exclude or confirm the diagnosis active Lyme disease. Assuming that at least a proportion of patients with presumed persistent Lyme disease indeed have chronic active B. burgdorferi infection, the sensitivity of ELISA is extremely low in this stage of disease. Indeed, in 205 patients responding to prolonged treatment, pretherapy ELISA was negative in 139 (68%). 17 This is not unexpected, since ELISA detects only antibodies against one single antigen, whereas expression of B. burgdorferi antigens is highly variable and is affected by early antibiotic treatment. 13 Although the Western Blot has been advocated as a second step to confirm positive ELISA tests 14-16, this strategy is not rational when aiming at improving the sensitivity of serology. Rather, Western Blot may detect antibodies to various B. burgdorferi antigens where the single antigen tested by ELISA is not expressed. Indeed, Western Blot was positive in 133 of 180 (74%) of patients with presumed persistent disease who had negative ELISAs. 17 Therefore, Western Blot should be evaluated as a single step diagnostic tool in late-stage Lyme disease. PCR may be considered as the gold standard for detection of persistent B. burgdorferi infection. Although few patients with persistent Lyme disease present with skin lesions or overt synovitis accessible for biopsy and PCR, we found positive PCR for B. burgdorferi in skin or synovial biopsies from patients with negative ELISA and a clinical diagnosis of persistent Lyme disease not responding to short courses of antibiotics. 3.3 Choice of antibiotics By many, ceftriaxone is considered as the gold standard of late-stage Lyme disease. However, intravenous ceftriaxone therapy is expensive, logistically cumbersome, and complicated by phlebitis and other adverse effects. Especially in cases of prolonged treatment, oral therapy would be desirable. Several studies have suggested that oral antibiotics may be as effective as ceftriaxone, even in late-stage disease. 65;67-69 For oral therapy of persistent infection doxycycline and some of the macrolides may be advantageous because of their excellent intracellular concentrations. Putative resistance to doxycycline after previous use may be of concern. Among the macrolides, roxitromycin has less favourable in vitro activity 70;71, whereas few data exist on prolonged treatment with azitromycin, which is licensed for short-term therapy only. Therefore, claritromycin is the macrolide of choice. Although its tolerability during prolonged treatment for possible Lyme disease is favourable (UMCN, data on file), no data on its efficacy for B. burgdorferi infection has been published. In contrast, the clinical experience with doxycycline is very large, both for short-term and prolonged therapy. The present study is designed to compare prolonged oral doxycyclin with combination therapy of claritromycin plus hydroxychloroquine after shortterm intravenous ceftriaxone for possible persistent Lyme disease. Lyme protocol version MAR 2010 Page 10

14 No consensus exists on the duration of treatment for this syndrome. Theoretical considerations and in vitro studies on persistent B. burgdorferi argue for treatment for more than 30 days. 61 Animal experiments showed persistence of viable B. burgdorferi in dogs after 30 days of doxycycline. 56 Therefore, a total of 14 weeks (98 days) of treatment has been chosen in the present study. Due to logistical and safety reasons, intravenous therapy will be limited to 14 days. 3.4 Assessments of outcome It has been shown that there is poor correlation between objective neuropsychological functioning and self-reported neuropsychological complaints in patients with diseases characterized by chronic fatigue. 72 Combined psychological testing and rating of complaints by questionnaires are required to assess the somatic and cognitive impairments in the subjects under study. Physical outcome a. Rating of physical outcome Self-report assessment of health-related quality of life (including pain and functional impairment) is measured with a scale that has previously been validated for patients with Lyme disease, the Medical Outcomes Study 36-item Short-form General Health Survey (SF- 36). 5 This scale has been used for large cohorts of patients with Lyme disease, including clinical trials. 5;73 The subjective feeling of fatigue is measured by the Fatigue Severity subscale of the Checklist Individual Strength (CIS) questionnaire. The CIS is a reliable instrument with good validity and sensitivity to change in patients with rheumatoid arthritis, fibromyalgia and chronic fatigue syndrome b. Accelerometer rating of physical activities Accelerometers are piezo-electric sensors which are worn by the patient to record physical activity over a period of several days. These devises, such as the Actometer (Medical Psychological Laboratory, Radboud University Nijmegen), are reliable and valid measures of human physical activity. 77 Objective Actometer records show modest correlations with selfreport questionnaires, in particular those requiring general subjective interpretations by the patient. 77 The combined use of self-report assessments of e.g. physical outcomes and daily activity levels with Actometer records enables to assess the multiple factors that are involved in the activity level of patients. c. Lancet Neurology score In order to evaluate neurological symptoms we will use a score representing standardized neurological examination. This clinical score was previously used to evaluate intravenous ceftriaxone treatment in Lyme neuroborreliosis. 78 This composite clinical score, from now on referred to as Lancet Neurology score, is based on standardized interview and clinical Lyme protocol version MAR 2010 Page 11

15 neurological examination. It consists of subjective complaints and objective findings. The subjective part consists of 6 items, each item is scored: 0=none, 1=mild, 2=severe; with a maximum score of 12. The objective part consists of 26 items, scored as above, with a maximum score of 52. Total maximum score is 64, and a score of 0 is the best possible score. The Lancet Neurology score will be assessed by doctors, trained by a neurologist. Cognitive outcome Rating of neuropsychological complaints and functioning Cognitive impairment and impaired concentration are one of the major problems reported by this patient population. This will be evaluated by standardized and validated neuropsychological self-report assessments, including Multifactorial Memory Questionnaire (MMQ). 79 In addition, a standard battery of neuropsychological tests will be used, including tests on memory, attention and motor function. Psychological outcome Rating of anxiety and depression The Hospital Anxiety and Depression Scale (HADS) will be used as standardized self-report questionnaires to measure psychological outcomes of anxiety and depression. 80 General outcome of functioning in daily life Rating of functioning in daily life The impact of the condition on different area s in daily life will be assessed by the Impact scale of the IRGL 81, a scale that measures the general impact of the condition on several area s in daily life (e.g. activities, work, leisure time, relationships) and that has previously been validated in various chronic somatic conditions, including patients with rheumatoid arthritis and fibromyalgia. 74;75;81;82 Furthermore, a modified version of the Fibromyalgia Impact Questionnaire (FIQ) will be used. This modified questionnaire has been validated for patients with persistent Lyme disease, and contains items to evaluate physical impairment, symptom severity, and global well-being. 83 Additional outcomes a. Assessment of moderators and mediators For the prediction of success of therapy and analysis of effectiveness in subgroups, cognitive-behavioral factors that have previously been shown to mediate or moderate outcomes of medical and multidisciplinary therapies in patients with chronic conditions 84;85 will be assessed by validated questionnaires of perceived helplessness (ZCL) 86 and negative outcome expectancies. 87. In addition, therapy compliance with the prescribed medication will be assessed by a daily logbook specifically developed for this study and adjusted from already existing compliance scales. 74 Lyme protocol version MAR 2010 Page 12

16 b. Assessment of cytokine profiles To study whether patients with persistent infection or poor response to antibiotic therapy may have abnormalities in their host defense against B. burgdorferi, the cytokine response of their peripheral blood mononuclear cells on B. burgdorferi-specific and aspecific stimuli will be studied as described earlier. 88 In addition, blood will be collected for DNA polymorphism analysis. 89 c. Economic evaluation This study investigates the potential efficiency (cost-utility analysis) of short provision of antibiotics (2 weeks standard) versus long provision of antibiotics (2 weeks standard + 12 weeks oral) from a societal perspective. Primary outcome measures for economic evaluation are costs and quality-adjusted life years (QALY s). The ratio cost per QALY gained (ICER) is computed and parameter uncertainty is determined using the bootstrap method. One-way sensitivity analyses on the range of extremes is applied. The economic evaluation is done along-side the clinical trial and consequently adheres to the earlier presented design. Cost analysis exists of two main parts. First, volumes of care are measured prospectively using patient-based diaries. Productivity losses for patients are estimated using the questionnaire Vragenlijst Ziekte en Werk. The friction cost method will be applied. Also travel time to therapy or outpatient clinic and related costs will be considered (included in patient based diary). The second part of the analysis determines prices using the Dutch guidelines for cost analyses. 90 For units of care/resources where no guideline or standard prices are available, real cost prices will be determined. For the overall quantification of health status as a single index (utilities) we will use the standard EQ-5D classification system developed by the EuroQol Group. 4. Study objective 4.1 Primary objective The primary goal of the study is to establish whether prolonged antibiotic treatment of patients diagnosed with presumed PLD (as endorsed by the international ILADS guidelines) leads to better patient outcome than short-term treatment as endorsed by the Dutch CBO guidelines. 63; Secondary objective Secondary objectives will be studied in an explorative way. The secondary goals include the effect of randomised treatment modalities on pain, functional impairment, psychological functioning, social behaviour, cognitive functioning, and safety. Moreover, cost-effectiveness will be determined by assessment of costs from societal perspective and quality-adjusted life Lyme protocol version MAR 2010 Page 13

17 years. 5. Study design This is a double blind, randomised, placebo-controlled trial of prolonged antibiotic treatment after intravenous ceftriaxone. All patients will initially receive open-label i.v. ceftriaxone in a home-care setting for two weeks, which is the standard of care for presumed or proven neuroborreliosis according to both guidelines. Then patients will be randomised to one of 3 treatment arms. Subsequently, blinded oral follow-on treatment will be given in 3 randomisation arms: I. oral doxycycline for 12 weeks; II. oral clarithromycin plus hydroxychloroquine for 12 weeks; or III. oral placebo for 12 weeks. The strategic choices leading to the design of a prospective, randomised, 3-arm study are complex. First, i.v. ceftriaxone followed by doxycycline is generally considered the gold standard therapy for complicated Lyme disease, e.g., by the NIH/NIAID Chronic Lyme study group. 5 However, its efficacy in PLD has never been proven in prospective studies. Whereas administration of ceftriaxone for longer than 2 weeks has been advocated, no published randomised studies on primary treatment of neuroborreliosis have addressed its use for more than 2 weeks, as underlined in the CBO guidelines. 63 Prolonged therapy with oral doxycycline has been associated with success in large, retrospective series of presumed PLD 17, but newer data from a retrospective series suggest than combined therapy with oral clarithromycin and hydroxychloroquine for at least 3 months may be at least as effective. 3 The present study involves assessment of management of patients with presumed PLD. In this study, clinical suspicion of PLD is defined as complaints of musculoskeletal pain, arthritis, neuralgia, neuropsychological or cognitive disorders, and persistent fatigue, that are (a) temporally related to an episode of erythema migrans or otherwise proven symptomatic Lyme disease (defined as within 4 months after erythema migrans or positive biopsy, PCR, culture, intrathecal antibodies), or (b) accompanied by a positive B. burgdorferi IgG or IgM immunoblot (as defined by strict criteria), regardless of prior ELISA IgG/IgM screening results. 42;91;92 Patients will be recruited from two hospitals, the Radboud University Nijmegen Medical Centre and the Sint Maartenskliniek Nijmegen. Screening will be done according standard clinical and laboratory protocols. After obtaining informed consent, baseline assessments include clinical, laboratory, microbiological and (neuro)psychological evaluation and objective assessment of physical activity, using an accelerometer. Study visits will be performed at baseline, week 2, 8 and 14 for safety evaluation. Efficacy evaluation will be performed at week 14 (end of treatment period, EOT), and at week 26 (12 Lyme protocol version MAR 2010 Page 14

18 weeks after EOT) and week 40 (end of study, EOS, 26 weeks after EOT), consisting of clinical and psychological assessment and accelerometer registration. After the last comprehensive outcome assessment at week 40, patients will be surveyed by post-study questionnaires on recurrent symptoms and additional health resource use at week 52. Compliance of antibiotic use will be assessed using MEMS caps (Medication Event Monitoring System) and diaries. 93;94 An independent safety and data review committee will monitor patients safety, and perform an interim assessment after EOT of 60 patients. Patients who have been randomized to ceftriaxone + placebo who experience ongoing symptoms after EOT will be offered antibiotic treatment by an independent physician off-study after the study period. 6. Informed consent & institutional review The benefits and risks of participation in the study will be explained to each subject prior to entering the study, and at that time written informed consent will be obtained. Patients of less than 18 years of age and those unable to give informed consent will not be entered into the study. Prior to initiation of the study, the study protocol and the informed consent form will have to be approved by the Institutional Review Board. The institution will have appropriate insurance covering liability to study subjects. 7. Criteria for subject selection 7.1 Inclusion criteria a. Males or non-pregnant, non-lactating females who are 18 years or older. b. Women of child-bearing potential must agree to use contraception methods other than oral contraceptives during the study therapy period, since failure of oral contraceptives due to long-term antibiotic use has been described and doxycycline might be teratogenic. c. Patients with presumed or proven PLD. In this study, clinical suspicion of PLD is defined as complaints of musculoskeletal pain, arthritis or arthralgia, neuralgia, neuropsychological or cognitive disorders, and persistent fatigue, that are: (a) temporally related to an episode of erythema migrans or otherwise proven symptomatic Lyme disease (defined as within 4 months after erythema migrans as assessed by a physician, or positive biopsy, PCR, culture, intrathecal B. burgdorferi antibodies), OR (b) accompanied by a positive B. burgdorferi IgG or IgM immunoblot (as defined by strict criteria in line with the European Union Concerted Action on Lyme Borreliosis (EUCALB); see appendix A), regardless of prior ELISA IgG/IgM screening results. d. Subjects must sign a written informed consent form. Lyme protocol version MAR 2010 Page 15

19 7.2 Exclusion criteria a. Subjects with a known history of allergy or intolerance to tetracyclines, macrolides, hydroxychloroquine or ceftriaxone. b. Subjects who have had more than 5 days of antimicrobial therapy with activity against B. burgdorferi within the previous 4 weeks. c. Subjects with a presumed diagnosis of neuroborreliosis (CSF pleiocytosis or intrathecal antibody production) for which intravenous antimicrobial therapy is required. d. Subjects with a known diagnosis of HIV-seropositivity or other immune disorders. (No HIV serologic testing is required for the study). e. Subjects with positive syphilis serology or signs of other spirochetal diseases. f. Subjects with moderate or severe liver disease defined as alkaline phosphatase, ALAT, or ASAT greater than 3 times upper limit of normal. g. Subjects who are receiving and cannot discontinue cisapride, astemizole, terfenadine, barbiturates, phenytoin, or carbamazepine (the concentrations of these drugs may increase during claritromycin therapy and/or lead to reduced availability of doxycycline). h. Subjects who are currently enrolled on other investigational drug trials or receiving investigational agents. i. Subjects who have been previously randomized into this study. j. Severe physical or psychiatric co-morbidity that interferes with participation in the study protocol, including previous medical diagnosis of rheumatic conditions, chronic fatigue syndrome or chronic pain conditions as well as insufficient command of the Dutch language. k. Co-morbidity that could (partially) account for the symptoms of the subject (e.g. vitamin B12 deficiency, anemia, hypothyroidism). 7.3 Randomisation procedure A subject will be considered eligible for study entry after the investigator has determined that the inclusion and exclusion criteria have been satisfied. Once it has been determined that a subject meets the entry criteria, the details of the study will be carefully discussed with the patient, who will be asked to read and sign the informed consent form. After obtaining informed consent, subjects meeting the inclusion and exclusion criteria will be randomised using a computer-generated randomisation list. Randomisation will be based on a minimization algorithm with the following variables: age, gender, duration of symptoms, and baseline SF-36 score. Treatment should remain blinded at all times. Section 13.4 describes the circumstances and procedures for breaking the blind for individual patients. Lyme protocol version MAR 2010 Page 16

20 8. Methodology 8.1 Pre-randomisation (baseline) evaluation The following must be performed preceding randomisation Complete medical history. Lyme disease history questionnaire Previous antibiotic therapy questionnaire. Standardized physical examination. Standardized neurological examination by the Lancet Neurology score. Microbiological serology, including B. burgdorferi ELISA and immunoblot (see Appendix A). Clinical laboratory: CRP, hemoglobin, hematocrit, MCV, leukocytes, leukocyte differentiation, platelets, calcium, albumin, glucose, creatinin, alkaline phosphatase, ASAT, ALAT, LDH, creatine kinase. Specific laboratory: anti-nuclear antibodies, thyroid stimulating hormone, iron, iron binding capacity, ferritin, vitamin B12 Biopsy for histopathology, B. burgdorferi culture, and PCR of any skin lesions suggestive of Lyme disease, if applicable. Synovial biopsy for B. burgdorferi culture and PCR of any arthritis of the knee or other large joints, if applicable. Blood and urine for cytokine stimulation tests and DNA polymorphisms. Baseline self-report assessments of physical, (neuro)psychological and general functioning and cognitive-behavioral factors, as described in section Neuropsychological and functional evaluation Evaluation of physical, psychological, cognitive and generic functioning is performed at baseline and at specified visits as described in section 8.3. Tests include: Medical Outcomes Study 36-item Short-form General Health Survey (SF-36) Fatigue subscale of Checklist Individual Strength (CIS) Body sensations questionnaire (BSQ) Hospital anxiety and depression score HADS (anxiety and depression) Impact of disease on daily life (IRGL) (general functioning) Fibromyalgia Impact Questionnaire (FIQ) (physical impairment, symptom severity, and global well-being) Actometer recording during 14 days (objective physical activity) Cognitive and memory complaints (MMQ) Illness Cognition Questionnaire ICQ (helplessness), negative outcome expectancies and therapy compliance scales McGill pain questionnaire (MPQ) Neuropsychological evaluation with the following tests will be performed by a Lyme protocol version MAR 2010 Page 17

21 psychologist: 15 words test, numberseries, trailmaking, stroop, symbol substitution, long-term 15 words test, Amsterdam short-term memory test (15 woorden test, cijferreeksen, trailmaking, stroop, symboolsubstitutie, lange termijn 15 woorden test en Amsterdamse korte termijn geheugentest). 8.3 Study visits Screening at baseline will be done according to standard clinical and laboratory protocols. If a patient shows signs of a neuroborreliosis, he/she will be referred to a neurologist. Patients with overt neuroborreliosis are identified by cerebrospinal fluid (CSF) examination; patients with CSF pleiocytosis will be excluded from the study. After obtaining informed consent, baseline assessments include clinical, laboratory, microbiological, physical and (neuro)psychological impairment evaluation and objective physical activity registration during 2 weeks, using an accelerometer. Study visits will be performed at baseline, week 2, 8 and 14 for safety evaluation. Efficacy evaluation will be performed at week 14 (end of treatment period, EOT), and at week 26 (12 weeks after EOT) and week 40 (end of study, EOS, 26 weeks after EOT), consisting of clinical, psychological, cognitive and general functioning assessment. Accelerometer registration will be performed at baseline, week 14 and 40. Questionnaires for economic evaluation will be performed at baseline, week 2, 14, 26 and 40. After the last comprehensive outcome assessment at week 40, patients will be surveyed by post-study questionnaires (SF-36) on recurrent symptoms and additional health resource use at week 52. Urine samples will be collected at baseline and week 14. Compliance of antibiotic use will be assessed using MEMS caps (Medication Event Monitoring System) and diaries. 93;94 Safety assessments (baseline, week 2, 8, 14) include: a. Medical history. b. Physical examination. c. Hemoglobin, hematocrit, leukocytes, platelets, glucose, creatinin, alkaline phosphatase, ALAT. Efficacy assessments (week 14, 26 and 40) include: Evaluation by questionnaires as described in section 8.2 Lancet Neurology score at week 14 Photography as well as biopsy for histopathology, B. burgdorferi culture, and PCR of any persistent skin lesions suggestive of Lyme disease, if applicable. Repeated synovial biopsy for B. burgdorferi culture and PCR of any persistent arthritis of the knee or other large joints, if clinically judged relevant for further management. Concomitant medication log. Drug accounting at week 14 and by MEMS caps and diaries. A flow chart of the study visits is shown in Appendix B. Lyme protocol version MAR 2010 Page 18

22 8.4 Accelerometers To record physical activity levels, accelerometers (Actometers) are worn by the patient during 2 weeks in combination with daily monitoring of activity levels. The Actometers will be worn at baseline, EOT (week 14) and EOS (week 40). 8.5 Follow up At week 52, twelve weeks after end of study (EOS), patients will be contacted by telephone and asked to complete a questionnaire addressing other medical diagnoses, additional antimicrobial therapy received, and persistent signs and symptoms. Patients who have been randomized to ceftriaxone + placebo experiencing ongoing symptoms at EOS will be offered antibiotic treatment by an independent physician off-study. 9. Ethical and regulatory considerations This study will be conducted in accordance with the ethical principles stated in the most recent version of the Declaration of Helsinki or the applicable International Conference on Harmonisation (ICH) guidelines on Good Clinical Practice. The study will start after approval by the Ethics Review Board (CMO) region Arnhem Nijmegen. Written informed consent will be obtained from the patients. 10. Study treatments 10.1 Ceftriaxone All patients will be treated with open-label i.v. ceftriaxone 2000 mg qd via a peripheral i.v. catheter. To monitor for side effects, patients will be admitted to the Sint Maartenskliniek for administration during day 1 and 2. Subsequent doses will be given in a home-care setting by specialized nurses, using an EasyPump administration system, prepared by the Sint Maartenskliniek Pharmacy Randomized study drug preparation Study drugs consist of doxycyclin 100 mg combined with a placebo b.i.d. or claritromycin 500 mg combined with hydroxychloroquine 200 mg b.i.d. for 12 weeks, or a 12 weeks' course of double placebo b.i.d.. Study drugs are to be taken twice daily after the meals. Study drugs and placebo will be prepared as capsules with identical appearance. The study drugs will be packaged in two different boxes containing 168 (+2 extra) capsules each (each treatment arm consists of 2 study-drugs b.i.d. for 84 days), labeled with the patient number and course number. Lyme protocol version MAR 2010 Page 19

23 10.3 Study drug supply and accounting After randomization, the patient will receive the study medication boxes for 12 weeks. The patient is required to bring the medication boxes at all visits, and drug utilization will be recorded by the investigator. Compliance of antibiotic use will be checked using MEMS caps (Medication Event Monitoring System) and daily questionnaires Unblinding of randomization The randomization assignment will be known to the study pharmacist only, and be available in a sealed envelope for each patient, stored at the principal investigator's office for emergency use. The code-break envelope should not be opened except in emergencies where knowledge of the treatment given is absolutely necessary for the further management of the patient. The reason for opening and date should be entered in the patient's case record form. If the code is broken under any circumstance, it will render the subject not eligible, as defined in section The investigators should contact the principal investigators before unblinding the treatment. An unblinded randomization list will not be made available to the investigators until the trial is completed and the analysis performed. 11. Concomitant medication Any other antibacterial drugs other than study medications are prohibited during the entire study period. In case of proven inter-current infections (e.g. urinary tract infection), specific antimicrobial therapy may be given for a maximum of 5 days. Indications should be discussed with the investigator, and efforts should be made to select an antimicrobial drug with no in vitro activity against B. burgdorferi. The following drugs are prohibited because of potential interaction with study drugs (a through f) or potential effects on efficacy of treatment (g through i): a. cisapride (Prepulsid ) b. astemizole (Hismanal ) c. terfenadine (Triludan ) d. barbiturates e. phenytoin (Diphantoïne, Epanutin ) f. carbamazepine (Tegretol, Carbimal ) g. prednisone h. recombinant cytokines or hematopoietic growth factors i. immunoglobulins If treatment with one of these drugs is required, the investigator should be contacted, study drug must be discontinued immediately, and the subject will be classified as therapy discontinuation, as defined in section Lyme protocol version MAR 2010 Page 20