A Randomised, Double Blind, Placebo Controlled Trial of Doxycycline in Lymphangioleiomyomatosis.
We will perform a 2 year double blind, placebo controlled trial of doxycycline in 40
patients with LAM. The main endpoints will be change in FEV1, other measures of efficacy,
safety and dose required to suppress MMP activity. After clinical evaluation, lung function,
shuttle walk, QoL assessment, blood tests, serum and urine MMPs (termed full assessment)
plus baseline chest X-ray patients will be randomised to doxycycline 50 mg od or placebo.
Patients will be assessed at 2 weeks for a safety screen, every 3 months for clinical
evaluation and spirometry and at 12 and 24 months for full assessment. At 3 months urine
zymography will be performed to see if MMPs are present in urine at the prescribed dose.
Doxycycline will be increased to 100 mg bd at three months after urine zymography has been
performed. To avoid withholding treatment from those who decline rapidly, patients who, on
two occasions, have either a fall from baseline FEV1 of 300 ml or fall in resting SaO2 of 3%
will be assessed by an independent expert (AET). Patients receiving placebo will be given
the option of doxycycline according to protocol. Those receiving doxycycline the option of
continuing in the study or withdrawal. These patients and those withdrawn due to recurrent
pneumothorax, increase in chylous effusion or bleeding angiomyolipoma will be included in a
composite safety endpoint and analysed on an intention to treat basis. Power calculations
based on retrospective cohorts(Johnson and Tattersfield 1999) show that 20 patients per
group will give 80% power to detect a 70 ml/year difference in FEV1 based an assumed SD for
a fall in FEV1 of 75 ml/year. The mean slope of regression lines for FEV1 and FVC, plus
change DLCO, shuttle walk distance and QoL in the doxycycline and placebo groups will be
compared by parametric or non-parametric analysis dependent on data, time to composite
safety endpoint by Caplan-Meier analysis, complications and adverse events by Chi Square
Lymphangioleiomyomatosis Lymphangioleiomyomatosis (LAM) is a disease of the lungs and
lymphatics, which can occur sporadically or in association with tuberous sclerosis complex
(TSC). The disease is rare, occurring in 1-2 / million of the population but in up to 40% of
women with TSC LAM almost exclusively affects women, generally developing before the
menopause. The disease is characterised by progressive pulmonary cystic change, recurrent
pneumothorax, chylous pleural collections and, in most cases, progressive respiratory
failure. Abdominal manifestations caused by obstruction and dilation of the axial lymphatics
include lymphadenopathy, cystic lymphatic masses (lymphangioleiomyomas), chylous ascites and
angiomyolipoma (a benign tumour). Survival in LAM is, 70 90% at 10 yrs, although this is
highly variable since long-term survivors have been described. Diagnosis is made by a
combination of clinical features and computed tomography scanning or, in cases of doubt,
lung biopsy. In patients with rapidly progressive disease, hormone treatment (predominantly
progesterone) has been used, although no firm evidence supports its use. Otherwise,
treatment is aimed at complications including pneumothorax, chylous collections and
extra-pulmonary manifestations. The only treatment for severe LAM is currently lung
transplantation(Johnson 2006). Recently identification of abnormalities in the tuberous
sclerosis complex (TSC) genes in sporadic and TSC associated LAM have identified
dysregulation of the mTOR pathway in LAM (Carsillo, Astrinidis et al. 2000; Sato, Seyama et
al. 2002) and have lead to clinical trials of mTOR inhibitors such as rapamycin in LAM and
TSC. At the time of writing these have not been reported but appear promising.
Background and preliminary data Cystic lung destruction is the hallmark of pulmonary LAM and
generally results in respiratory failure over a variable period of time(Johnson 2006).
Over-activity of proteases including elastase, trypsin and the matrix metalloproteinases
(MMPs) is responsible for parenchymal destruction in emphysema and other lung diseases. The
MMPs are a family of zinc dependent proteolytic enzymes with proteolytic activities against
extra-cellular matrix proteins. The MMPs are overexpressed in inflammatory and neoplastic
diseases where in addition to processing extra-cellular matrix they also have roles in
metastasis, angiogenesis, growth factor activation and inactivation(Stamenkovic 2003). MMPs
-1, -2, and -9 are involved in the sequential digestion of collagen and gelatin and are
strongly expressed in the walls of cysts where it is thought they contribute to parenchymal
destruction in LAM(Matsui, Takeda et al. 2000). Further, MMPs can be detected by gelatin
zymography in the urine of patients with LAM but not controls. Doxycycline is a tetracycline
antibiotic in common clinical use. In addition to its antimicrobial action it inhibits the
synthesis and activity of several MMPs and inhibits proliferation of a range of cell types
including arterial smooth muscle, cancer cells and cancer model systems(Bendeck, Conte et
al. 2002; Duivenvoorden, Popovic et al. 2002; Onoda, Ono et al. 2004). In preliminary
experiments we have demonstrated that primary LAM derived and angiomyolipoma cells produce
MMP-2 and -7 and that MMP expression and proliferation in these cells in inhibited by
Preliminary clinical data of doxycycline in LAM In a preliminary open label study of
doxycycline (50 - 100 mg qds) in 10 patients with LAM, doxycycline improved 6 minute walk
distance) and Borg dyspnoea score (Glassberg et al. Data presented at the LAM Foundation
International Research Conference, Cincinnati Ohio 2006). In a single case report, Moses et
al. observed an improvement in FEV1 and oxygenation during exercise in a patient with LAM
treated with 100mg doxycycline daily (Moses, Harper et al. 2006). In both of these cases
MMP-2 and -9 were initially present in the patient's urine and was undetectable after
treatment with doxycycline. In a study of 14 patients, doxycycline 100mg bd is known to give
a mean plasma concentration of 4.41 µg/ml (range 1.9-9.4 µg/ml)(Prall, Longo et al. 2002).
Specific issues in orphan disease clinical trials and rationale for trial design Studying
orphan diseases presents specific challenges, specifically the limited number of patients
available, wide geographic distribution and low priority for funding due to the perceived
poor economic benefit. Patients are well informed about potential developments due to
patient groups and internet based information(Tattersfield and Glassberg 2006) and may
obtain potential treatments 'off label' making definitive research studies impossible. As
LAM is rare, cohorts drawn from a wide area are required for studies to achieve adequate
power. We have 15 years experience of LAM research and from our UK LAM database estimate
there are approximately 120 patients in the UK. We are currently performing an open label
study of Sirolimus in LAM and tuberous sclerosis (TESSTAL, a Study of The Efficacy and
Safety of Sirolimus (Rapamycin) Therapy for Renal Angiomyolipomas in Patients with Tuberous
Sclerosis Complex And Sporadic Lymhangioleiomyomatosis). This included six patients with LAM
whom we found to be well motivated and will travel long distances for study visits
(including Cornwall, Kent and Perth). Despite the known adverse effects of rapamycin six of
eight eligible (i.e. with LAM and angiomyolipoma) patients invited took part in the study.
The current study protocol has been designed as a simple protocol which is both inclusive of
most patients with LAM and has a follow up which is similar to routine clinical care which
we hope will facilitate recruitment. In addition the protocol is less demanding than that of
the TESSTAL study. As most patients are known to us via our database or clinical contacts we
expect recruitment to be complete within six months.
Designing a definitive study is difficult without having an estimate of the likely size of
effect, if this is large, as suggested by the one case report(Moses, Harper et al. 2006) a
small number of patients are needed. If it is small as seems more likely a priori, a larger
number of patients are needed requiring European collaboration and considerably greater
funding. We therefore designed a pragmatic pilot study using a single geographic population
with a simple, inexpensive protocol which will serve several functions, specifically to: (1)
determine the optimum dose of doxycycline needed to suppress MMP production. (2) define the
safety profile of doxycycline in LAM. (3) provide evidence of efficacy and size of effect.
(4) provide data to help optimise the design and logistics of future trials.
2 Study aims/objectives
Hypothesis Doxycycline will prevent matrix metalloproteinase dependent tissue destruction in
lymphangioleiomyomatosis (LAM) thus preserving lung function, exercise capacity and quality
We will perform a randomised placebo controlled trial of doxycycline on rate of decline of
FEV1 over two years against matched placebo. This study will:
1. determine the optimum dose of doxycycline needed to suppress MMP production.
2. define the safety profile of doxycycline in LAM.
3. provide evidence for efficacy and size of effect.
4. provide data to help optimise the design and logistics of future trials.
With the data obtained we will be in a strong position to apply for European funding for a
European wide trial should this still be required.
3 Investigational plan
Patient population and recruitment
Forty patients with either sporadic LAM or TSC-LAM will be recruited from the UK LAM
database, physician referrals, LAM Action (a patients group for women with LAM) and the
Tuberous Sclerosis Association. Patients will be contacted by mail by the principal
investigator. Potential participants will receive a preliminary information sheet and
response sheet for return by prepaid post. Those that express a potential interest in
participating will be offered a face to face interview with one of the study doctors to
assess eligibility, answer questions, obtain details of all physicians involved in their
care and obtain written consent prior to enrolment. We anticipate recruitment will be
complete within 12 months of starting the study.
Allocation: Randomized, Endpoint Classification: Safety/Efficacy Study, Intervention Model: Parallel Assignment, Masking: Double Blind (Subject, Caregiver, Investigator), Primary Purpose: Treatment
Mean rate of change of FEV1 over 24 months on doxycycline compared with placebo.
Simon R Johnson, DM FRCP
University of Nottingham
United Kingdom: Medicines and Healthcare Products Regulatory Agency