62574

CONFIDENTIAL

National Institute for Health and Clinical Excellence Page 1 of 51

Premeeting briefing Pegloticase for treating severe debilitating chronic tophaceous gout

Issue date: November 2012

NATIONAL INSTITUTE FOR HEALTH AND CLINICAL EXCELLENCE

Premeeting briefing

Pegloticase for treating severe debilitating chronic tophaceous gout

This premeeting briefing is a summary of:

the evidence and views submitted by the manufacturer, the consultees and their nominated clinical specialists and patient experts and

the Evidence Review Group (ERG) report.

It highlights key issues for discussion at the first Appraisal Committee meeting and should be read with the full supporting documents for this appraisal. Please note that this document is a summary of the information available before the manufacturer has checked the ERG report for factual inaccuracies.

Key issues for consideration

Clinical effectiveness

The manufacturers submission presented a treatment pathway based on

British Society for Rheumatology (BSR) guidelines and the European

League Against Rheumatism (EULAR) guidelines. It stated that, in patients

whose serum uric acid has not normalised and whose signs and symptoms

are inadequately controlled with xanthine oxidase inhibitors at the

maximum medically appropriate dose, or for whom these medicines are

contraindicated, current standard care is best supportive care consisting of

non-steroidal anti-inflammatories (NSAIDs), colchicine and corticosteroids.

The Evidence Review Group (ERG) and the its clinical advisers agreed

with the manufacturer that pegloticase is likely to be implemented in clinical

practice only after the first 2 lines of treatment recommended in the BSR

and EULAR guidelines (xanthine oxidase inhibitors and uricosurics) have

been exhausted. Does the Committee agree that best supportive care is

the only available treatment in clinical practice for the patient population

likely to be covered in the UK marketing authorisation for pegloticase?

CONFIDENTIAL




Given that:

The professional groups highlighted that allopurinol intolerance is often

wrongly diagnosed in clinical practice, without appropriate titration of

dosage or any attempt of desensitisation.

Patients in the trial population described in the manufacturers

submission either couldnt have allopurinol because it was

contraindicated or their uric acid levels had not normalised despite 3 or

more months of treatment with the maximum medically appropriate dose

determined by the treating clinician.

The ERG stated that it was not clear from the manufacturers submission

whether desensitisation to allopurinol hypersensitivity had been

attempted before trial entry.

Does the Committee consider that the results of these trials can be

generalised to clinical practice in the NHS in England and Wales?

Clinical effectiveness evidence in the manufacturers submission was

based predominantly on the findings from simple pooled analyses of

primary and secondary efficacy data from 2 phase III trials. The

manufacturer stated that because these trials were replicates, a meta-

analysis of the combined estimate of effects would not be appropriate. But

the ERG pointed out that simple pooling of data may yield counterintuitive

results. Does the Committee consider simple pooling of outcome data

acceptable?

The manufacturers submission stated that that there were no unexpected

imbalances in drop-outs between groups in the trials. But the ERG

highlighted that in the 2 key trials a higher proportion of patients in the

pegloticase every 2 weeks arms had dropped out at the end of the study

(30% in trial C0405 and 28% in trial C0406) than in the placebo arms (5%

in trial C0405 and 13% in trial C0406). The ERG stated that the

manufacturer did not explain or adjust for the imbalances in these the drop-

out rates, particularly in the patients who dropped out because of adverse

CONFIDENTIAL




events in both trials. Does the Committee think that this dropout may have

biased the results?

According to the manufacturers submission, one of the inclusion criteria

used to define chronic refractory gout in the phase III trials was gouty

arthritis, while in the published results from the trials (Sundy et al., 2011),

it was referred to as gouty arthropathy . At clarification stage the

manufacturer stated that gouty arthritis was used as an eligibility criterion,

but that the term gouty arthritis was not defined. It was left to the clinical

judgement of the investigator to decide whether or not the patient had this

condition. The ERG stated that it was unclear how gouty arthritis was

defined and applied as one of the inclusion criteria in the key clinical trials.

The marketing authorisation for pegloticase is for the treatment of severe

debilitating chronic tophaceous gout in adult patients who may also have

erosive joint involvement. Does the Committee consider that these terms

(gouty arthritis, gouty arthropathy or erosive joint involvement) refer to the

same condition?

In one treatment arm of the phase III trials patients were given pegloticase

8 mg every 2 weeks. The ERG stated that the impact of repeated courses

of pegloticase on uric acid levels, secondary outcomes, immunogenicity

and adverse events was not clear from the clinical effectiveness evidence

in the manufacturers submission. What is the Committees view on the

possible loss of efficacy and increased risk of adverse events because of

pegloticases potential immunogenicity?

Cost-effectiveness

In the manufacturers economic model, patients receiving pegloticase are

separated into responder, non-responder and non-completer groups.

The manufacturers submission stated that the primary reason for patients

becoming non-completers was adverse events and in clinical practice this

would occur within the first month of treatment. Data from the trials on the

proportion of non-completers who had withdrawn from pegloticase

CONFIDENTIAL




treatment by one month and the time to treatment cessation for non-

completers were unavailable. The ERG stated that it was uncertain how

patients in the non-completer groups would be identified in clinical

practice. What is the Committees view on this group in light of UK clinical

practice?

In the manufacturers health economic model it is assumed that the

maximum duration of pegloticase treatment would be 6 months. But data

from an open-label extension study of the patients included in the phase III

trials suggest average treatment duration of ************** in pegloticase

responders. Does the Committee accept that pegloticase would not be

used for more than 6 months in UK clinical practice?

The economic model presented in the manufacturers submission assumes

that the benefits of pegloticase treatment to responders during the 6-month

treatment period can be maintained in the long-term by allopurinol or

febuxostat treatment. The manufacturer did not provide direct clinical

evidence to support a maintained treatment effect using these treatments.

The ERG considered that this was an area of substantial uncertainty

because there was no direct evidence about effectiveness of maintenance

treatment. What is the Committees view on this uncertainty?

In the manufacturers model, persistence with maintenance treatment

(allopurinol or febuxostat) after pegloticase treatment in responders was

extrapolated based on indirect evidence. The ERG noted that 34% of the

incremental quality-adjusted life year (QALY) gains were accrued more

than 10 years after pegloticase treatment and the incremental cost-

effectiveness ratio (ICER) was sensitive to the discontinuation rate of

maintenance treatment. Does the Committee think the assumption about

persistence with maintenance treatment in the model is justifiable?

The ERG noted that treatment effectiveness is captured in the model by

patient-related outcomes including serum uric acid, frequency of gouty

flares and tophi resolution. Does the Committee think that a lower serum

CONFIDENTIAL




uric acid level could result in additional utility benefits over and above that

associated with reduced gout flares and tophi resolution?

The manufacturers model made assumptions about rheumatology visits,

accident and emergency visits and hospital admissions associated with

higher levels of serum uric acid. Does the Committee consider that

resource use would be higher with higher serum uric acid levels over and

above that associated with gout flares?

The manufacturers model captured treatment effectiveness by changes in

4 patient-related outcomes: serum uric acid, quality of life, frequency of

flares and tophi resolution. The ERG noted that the clinical continuation

rule in the manufacturers model differed from the definition of response

applied in the phase III trials, and that additional analyses were performed

on the trial data to inform its health economic model. Also, primary end

points in the trials were based on plasma uric acid level but while in the

manufacturers health economic model the end points were based on

serum uric acid level. What is the committees view on the data used in the

manufacturers modelling?

1 Background: clinical need and practice

1.1 Gout is a chronic, progressive and destructive condition that

causes acute, intermittent and painful attacks of arthritis in the

joints of the foot (especially the big toe), knee, hand and wrist. It is

caused by elevated uric acid levels in the blood (hyperuricaemia)

which leads to formation of monosodium urate crystals in and

around joints. Release of these crystals into the joint space

initiates an acute inflammatory response. With persistently high

levels of uric acid in the blood, gout may progress from acute

episodic attacks to a disabling, chronic, deforming arthropathy,

with destructive deposits of urate crystals (tophi) in bones, joints,

subcutaneous tissue and other organs. Renal damage may occur

because of urate crystal deposition and urinary tract stones

CONFIDENTIAL




composed entirely or partly of uric acid crystals. Severe gout with

recurrent flares and tophi has a significant impact on the patient

leading to pain, fatigue, progressive loss of mobility and function

and a worsening of quality of life.

1.2 The estimated prevalence of gout between 2000 and 2005 in the

UK was 1.4%. It affects more men than women. Prevalence

significantly increases with age, and was reported to be 7.3% in

men over 65, with the highest prevalence estimates in older men.

Both the incidence and prevalence of gout are increasing. This is

likely to be because of an ageing population, increasing levels of

obesity, and dietary changes. In line with the anticipated

marketing authorisation the manufacturer has estimated that ****

people in England and Wales will be eligible for treatment with

pegloticase.

1.3 Gout is treated by diet and lifestyle changes and drug treatments

tailored to risk factors and clinical stage. The aim of treatment is

to manage acute flares and, in patients with chronic gout, reduce

recurrent flares and tophi by reducing serum uric acid

concentration. Acute gout flare is self-limiting and generally lasts

for 7 to 10 days if untreated. Treatment relieves pain and speeds

recovery. Acute flares can last longer if the person has severe

gout or if the flare is inadequately managed. Drugs used to treat

acute gout flares include NSAIDs, colchicine and corticosteroids.

Practical measures such as applying ice and diet and lifestyle

changes can also help reduce the number and severity of flares.

Treatment to lower uric acid is recommended for patients with

recurrent acute episodes. The aim is to reduce serum levels to

below saturation point so the urate crystals dissolve and no new

crystals can form. The 2006 EULAR guideline recommends

reducing serum uric acid to below 360 micromol/litre while the

CONFIDENTIAL




British Society for Rheumatology and the British Health

Professionals in Rheumatology guideline for the management of

gout (BSR guideline) recommends a stricter target of

300 micromol/litre.

1.4 Two classes of uric acid lowering drugs are available: xanthine

oxidase inhibitors (allopurinol and febuxostat) and uricosurics

drugs (sulphinpyrazone, probenecid and benzbromarone). Both

the BSR and EULAR guidelines recommend allopurinol as a first-

line treatment. Febuxostat for the management of hyperuricaemia

in people with gout (NICE technology appraisal guidance 164)

recommends febuxostat as an option for managing chronic

hyperuricaemia in gout only for people who are intolerant of

allopurinol or for whom allopurinol is contraindicated. Both the

BSR and EULAR guidelines recommend uricosuric drugs

(sulphinpyrazone, benzbromarone [BSR guideline]; probenecid

and sulphinpyrazone [EULAR guideline]) second line as an

alternative to allopurinol. There are no data to suggest that

uricosuric drugs are effective in reducing tophi in patients with

severe tophaceous gout. Currently standard care for patients

whose serum uric acid has failed to normalise and whose signs

and symptoms are inadequately controlled with xanthine oxidase

inhibitors at the maximum medically appropriate dose, or for

whom these medicines are contraindicated, is best supportive

care consisting of treatment with NSAIDs, colchicine and

corticosteroids. Best supportive care manages acute flares but

does not provide the patient with effective long-term treatment to

reduce uric acid levels and improve clinical outcomes such as

resolution of tophi.

CONFIDENTIAL




2 The technology

2.1 Pegloticase (Krystexxa, Savient), is a polyethylene glycol

conjugate of recombinant uricase (urate oxidase). It catalyses the

oxidation of uric acid to allantoin, which is water soluble and can

be excreted in urine resulting lowered serum uric acid.

2.2 Pegloticase does not currently have a UK marketing authorisation.

It has received a positive opinion from the Committee for

Medicinal Products for Human Use (CHMP) recommending the

granting of a marketing authorisation for: The treatment of severe

debilitating chronic tophaceous gout in adult patients who may

also have erosive joint involvement and who have failed to

normalise serum uric acid with xanthine oxidase inhibitors at the

maximum medically appropriate dose, or for whom these

medicines are contraindicated.

2.3 The recommended dose of pegloticase is 8 mg given as an

intravenous infusion every 2 weeks. Before starting treatment with

pegloticase, patients should stop taking oral urate-lowering

medication. To minimise the risk of infusion-related reactions,

patients should receive pre-medication such as an antihistamine

the evening before, and again approximately 30 minutes before

the infusion. They should also take paracetamol and a

corticosteroid immediately before each infusion. The serum uric

acid level should be monitored before each infusion. Pegloticase

should be stopped if the patient has 2 consecutive serum uric acid

levels above 6 mg/dl (360 micromol/litre). Pegloticases optimal

treatment duration has not been established. Treatment duration

should be based on maintenance of response (serum uric acid

levels below 6.0 mg/dl) and clinical judgment.

CONFIDENTIAL




2.4 The summary of product characteristics lists the following most

common adverse reactions for pegloticase: infusion-related

reactions, gout flare, nausea, dermatitis, urticaria, pruritus, skin

irritation, dry skin, anaphylaxis, influenza-like illness, joint

swelling, vomiting, hyperglycaemia, hyperkalaemia, cellulitis and

haemolysis. For full details of adverse reactions and

contraindications, see the draft summary of product

characteristics (appendix 1of the manufacturers submission).

2.5 The acquisition cost of pegloticase has not been finalised. The

anticipated price for a single vial containing 8 mg of pegloticase

concentrate for 1 infusion is 1770 (excluding VAT). The CHMPs

summary of positive opinion states: The decision to treat with

[pegloticase] should be based on an ongoing assessment of the

benefits and risks for the individual patient. The average cost of a

course of 6 months treatment, as presented in the manufacturers

submission, is estimated to be 23,010. Costs may vary in

different settings because of negotiated procurement discounts.

3 Remit and decision problem(s)

3.1 The remit from the Department of Health for this appraisal was:

To appraise the clinical and cost effectiveness of pegloticase

within its licensed indication for the treatment of hyperuricaemia in

people with symptomatic gout whose disease is refractory to

conventional urate-lowering therapy, or in whom conventional

urate-lowering therapy is contraindicated or not tolerated.

CONFIDENTIAL




Final scope issued by NICE Decision problem addressed in the submission

Population Adults with hyperuricaemia and symptomatic gout whose disease is refractory to conventional urate-lowering therapy or in whom conventional urate-lowering therapy is contraindicated or not tolerated.

Adult patients with severe debilitating chronic tophaceous gout who have failed to normalise serum uric acid with xanthine oxidase inhibitors at the maximum medically appropriate dose, or for whom these medicines are contraindicated.

3.2 The population in the decision problem addressed in the

submission is more restrictive than the final scope. The

manufacturer highlighted that the likely anticipated indication was

a small subset of the refractory chronic gout population defined in

the scope. The CHMP positive opinion has a further restrictive

clause that treatment should be in adult patients who may also

have erosive joint involvement. But this clause isnt included in

the description of the population addressed in the submission.

The ERG did not have any major concerns with the

generalisability of trial populations to UK clinical practice.

However, it noted that the terms gouty arthritis and gouty

arthropathy had not been defined in the trials and also that it was

not clear whether these terms indicate erosive joint involvement.

The ERG also noted that in the phase III trials patients either

couldnt have allopurinol because it was contraindicated or their

uric acid levels had failed to normalise despite 3 or more months

of treatment with the maximum medically appropriate dose

(determined by the treating doctor). The ERG noted that it was

unclear from the manufacturers submission whether

desensitisation to allopurinol hypersensitivity had been attempted

before trial entry because no data were available on the maximum

doses or duration of previous urate-lowering treatments.


CONFIDENTIAL




3.3 The manufacturer submission stated that, while the phase III trials

were 6 months long, the optimal treatment course has not been

established. It stated that treatment duration should be based on

a maintained response (serum uric acid below 360 micromol/litre)

and clinical judgement. The manufacturer estimated that the

average length of a course of pegloticase treatment for patients

whose condition responded to treatment (and who maintain serum

uric acid levels below 360 micromol/litre was 6 months. The

treatment duration for patients whose condition did not respond to

pegloticase treatment was estimated as 2 months. The average

duration of treatment across all patients receiving pegloticase was

estimated by the manufacturer to be 4 months. The manufacturer

also reported that no information was available about the

anticipated number of repeat courses and treatments that may be

needed by patients receiving pegloticase and that the intervals

between treatments would be patient dependent. Further details

received at clarification stage on time to treatment stopping

across all pegloticase trials suggests average treatment duration

of around ******* in pegloticase responders.

Intervention Pegloticase


Comparators The standard comparators to be considered include:

Best supportive care.

Febuxostat for adults who are intolerant to allopurinol or for whom allopurinol is contraindicated.

Best supportive care

CONFIDENTIAL




3.4 The manufacturer stated that febuxostat is not an appropriate

comparator because patients whose serum uric acid normalises

with febuxostat (which is a xanthine oxidase inhibitor) will be

outside the licensed indication for pegloticase. The ERG pointed

out that the BSR and EULAR guidelines for the management of

gout both recommend using uricosuric drugs as a second-line

treatment after treatment with allopurinol. The ERG agreed that

uricosurics drugs are likely to be tried before pegloticase and

therefore best supportive care is an appropriate comparator for

the population likely to receive pegloticase in clinical practice.


Outcomes The outcome measures to be considered include:

serum urate levels

gout flares

reduction in tophus

pain

tender and swollen joint

physical function

adverse effects of treatment

health-related quality of life.

Serum and plasma urate levels

Gout flares

Tophus resolution

Pain

Tender and swollen joint count

Physical function

Adverse effects of treatment

Health related quality of life (Health Assessment Questionnaire [HAQ], HAQ disability index and SF-36)

3.5 The manufacturers submission included both serum and plasma

urate levels as the primary outcomes. Plasma uric acid was

preferred over serum uric acid as the primary outcome. This is

because during serum processing pegloticase could cause

enzymatic degradation of the uric acid in blood samples left at

room temperature, resulting in inaccurately low uric acid levels.

This does not happen when plasma is used to measure plasma

uric acid levels because the processing conditions limit the

enzymatic activity of pegloticase. The methods used to measure

plasma and serum uric acid levels were described by the

manufacturer as validated methods available in clinical

CONFIDENTIAL




laboratories. The manufacturer tested the observed plasma and

serum uric acid for agreement using the kappa coefficient and

obtained a value of 0.74 (p

CONFIDENTIAL




published article (Sundy et al. 2011) presented the results of

2 phase III trials of pegloticase: C0405 and C0406. The abstracts

presented data on some aspects of these trials and an open-label

extension study C0407. Data from the abstracts were not included

in the submission. The clinical data presented in the submission

were from the published article, an internal report on the

integrated summary of efficacy of these trials as well as an

unpublished clinical study report of open-label extension safety

study (C0407).

4.2 C0405 and C0406 were identical, randomised, double-blind,

placebo-controlled, multicentre, 3-armed trials of 6 months

duration conducted in the USA and Canada (C0405) and the USA

and Mexico (C0406). The trial patients were 18 years or older,

had a baseline serum uric acid level of at least 480 micromol/litre

and at least 1 of the following criteria: 3 or more self-reported gout

flares during the previous 18 months; 1 or more tophi; and gouty

arthropathy (clinical or radiographic evidence of joint damage due

to gout). Patients also had to either have a contraindication to

treatment with allopurinol or their uric acid levels must have failed

to normalise despite 3 or more months of treatment with the

maximum medically appropriate allopurinol dose (determined by

the treating doctor). Exclusion criteria were glucose-6-phosphate

dehydrogenase (G6PD) deficiency, previous treatment with a

uricase-containing drug, pregnancy, unstable angina, uncontrolled

hypertension (blood pressure higher than 150/95 mm Hg) or

cardiac arrhythmia, uncompensated congestive heart failure, renal

dialysis, or solid organ transplant. Intravenous pegloticase 8 mg

every 2 weeks and intravenous pegloticase 8 mg every 4 weeks

were compared with placebo. During the trial all patients received

prophylaxis against gout flare with colchicine or an NSAID, which

started 1 week before the first infusion of pegloticase and

CONFIDENTIAL




continued throughout the trial. Patients also received prophylaxis

against infusion reactions with oral fexofenadine 60 mg the

evening before each infusion, 60 mg oral fexofenadine and 1 g

oral paracetamol the morning of the infusion and intravenous

hydrocortisone 200 mg immediately before each infusion.

Although the trials included 2 pegloticase dosing regimens (8 mg

every 2 weeks or every 4 weeks) only the results of the patients

receiving the anticipated licensed dose (pegloticase 8 mg every

2 weeks) were included for the estimate of efficacy in the

submission.

4.3 The primary outcome was the proportion of patients whose

plasma uric acid responded to treatment in each pegloticase

treatment group compared with placebo. In the trials, a responder

was defined as a patient with a plasma uric acid level below

360 micromol/litre for at least 80% of the time during months 3

and 6. Plasma uric acid level was measured at baseline and 2

and 24 hours after the first infusion, before each biweekly infusion

and at 5 additional prespecified time points in month 3 and month

6 (2 hours, 1 day and 7 days after the week 9 and week 21

infusions and 2 hours and 7 days after the week 11 and week 23

infusions).

4.4 Secondary end points were assessed at baseline, week 13,

week 19 and week 25 and included resolution of tophi (defined as

a 100% decrease in area of at least 1 prespecified target tophus

without progression or appearance of a new tophus), reduction in

gout flares (frequency and incidence) and the number of flares per

patient during months 13 and 46, improvement in tender and

swollen joint count, improvement in quality of life (SF-36) and

improvement in functional status (Health assessment

questionnaire [HAQ] disability index and the HAQ pain scale).

CONFIDENTIAL




4.5 A total of 225 patients (109 in C0405 and 116 in C0406) were

randomised to the 3 trial groups (pegloticase every 2 weeks or

every 4 weeks, or placebo) in a 2:2:1 ratio. All urate-lowering,

clinical efficacy, and tolerability analyses (except deaths) were

carried out on a modified intent-to-treat population (n=212; 104 in

C0405 and 108 in C0406) comprising all randomised patients who

received at least 1 infusion. Baseline characteristics were similar

across the trials and treatment groups except for chronic kidney

disease (defined as a creatinine clearance below 60 ml/min),

which was markedly lower in the placebo group in C0406 (13%)

compared with the intervention groups in both the trials and the

placebo group in C0405 (2833%). Metabolic and cardiovascular

disorders were also common. More than 80% of trial patients had

cardiovascular comorbidities (see table 6.5, page 40 of the

manufacturers submission).

4.6 Plasma uric acid normalised within 24 hours of the first infusion

for all patients receiving pegloticase. But over time some patients

lost the urate-lowering response whereas others maintained a uric

acid level under 360 micromol/litre throughout the trials. The

number of responders in the treatment arms receiving 8 mg

pegloticase every 2 weeks (responders defined as patients whose

plasma uric acid level is under 360 micromol/litre for at least 80%

of the time during months 3 and 6) was 20/43 (47%) in C0405 and

16/42 (38%) in C0406. The primary end point was not achieved in

any placebo-treated patient and the results demonstrated a

significant difference for both trials (p

CONFIDENTIAL




4.7 The results of the secondary outcomes are presented as pooled

data from the pegloticase every 2 weeks arms of the 2 trials.

Resolution of tophi was measured in patients with tophi at

baseline. The proportion with complete resolution at month 6 was

statistically significantly higher in the pegloticase treatment group

than the placebo group (21/52 [40%] and 2/27 [7%] respectively,

p=0.002). The incidence of gout flares increased in the

pegloticase treatment group compared with the placebo treatment

group during the first 3 months of treatment (64/85 [75%] and

23/43 [53%] respectively, p=0.02). But with continued treatment,

statistically significant reductions in the incidence of flares during

months 4 to 6 were seen in the pegloticase treatment group

compared with placebo (28/69 [41%] and 29/43 [67%]

respectively, p=0.007). Similarly, during the first 3 months of

treatment, flare frequency (number of flares per patient) was

statistically significantly higher (p=0.001) in the pegloticase

treatment group (mean 2.3, standard deviation [SD] 2.1) than the

placebo group (mean 1.2, SD 1.5). During months 4 to 6 there

were fewer flares per person in the pegloticase group (mean 0.8,

SD 1.2) than in the placebo group (mean 1.3, SD 1.5) but this was

not statistically significant (p=0.06).

4.8 Patients in the pegloticase group also experienced reductions in

the number of tender and swollen joints at final visit from baseline

compared with those in the placebo group. The reduction in the

average number of tender joints was statistically significant in the

pegloticase group compared with the placebo group (7.4 and

1.2 respectively, p=0.01). There was also a greater reduction in

the number of swollen joints in the pegloticase group than in the

placebo group but this was not statistically significant (5.5 and

2.6 respectively, p=0.18).

CONFIDENTIAL




4.9 Patient-reported outcomes were also reported to be improved with

pegloticase. Statistically significant changes from baseline in HAQ

disability index scores and SF-36 physical component summary

(PCS) scores were observed, with the changes meeting or

exceeding the established minimum clinically important

differences.

4.10 HAQ disability index was measured by administering a survey

consisting of 20 questions about various physical activities

including activities of daily living. The patients were asked to

score their functional ability from 0 (no difficulty) to 3 (unable to do

without help or use of aids). The individual scores were averaged

to obtain a final score between 0 and 3. The mean change in the

HAQ disability index scores from 13 months to the final visit in

the pegloticase group was 0.22 (SD 0.64) compared with 0.02

(SD 0.41) for the placebo group. This was statistically significant

(p=0.01). The minimum clinically important difference reported in

the literature on the HAQ disability index for inflammatory arthritis

is 0.22.

4.11 The HAQ pain score was measured on a visual analogue scale

from 0 to 100 mm. The pain score at baseline was lower in the

pegloticase group (44.2 [SD 27.7]) than the placebo group (53.9

[SD 28.1]) although the difference was not statistically significant

(p=0.07). The reduction in pain score at final visit in the

pegloticase group was 11.4 (SD 33.8), which was greater than

the minimum clinically important difference for inflammatory

arthritis reported in the literature (0.10) as well statistically

significantly better (p=0.03) than the change in the placebo group

(1.4 [SD 30.0]).

CONFIDENTIAL




4.12 All patients in the pegloticase treatment group with data at

end point also reported statistically significant improvements in

their SF-36 PCS scores compared with placebo (p=0.01).

4.13 Secondary outcomes were also presented for plasma uric acid

responders compared with non-responders. Results for the

pooled analysis are in table 1. Compared with non-responders,

the responder group had a higher proportion of patients with

complete tophus response, a reduced incidence of flares during

months 1 to 3, a numerically higher reduction in the number of

swollen or tender joints and greater improvement in mean HAQ

pain and SF-36 PCS scores.

Table 1 Secondary endpoints responders compared with non-responders (pooled analysis)

Pegloticase 8 mg every 2 weeks Placebo (n=29)

Responders

(n=25 )

Non-responders

(n=37)

Resolution of tophi (as defined in section 4.4)

No. of patients/no.

evaluable patients

at final visit (%)

13/21 (61.9) 8/31 (25.8) 2/27 (7.4)

Flare incidence

Months 13

n/N (%) ************ ************ 23/43 (53.5)

Months 46

n/N (%) ************ ************ 29/43 (67.4)

Flare frequency per patient

Months 13

n ** ** 43

Mean (SD) ********** ********** 1.2 (1.62)

Months 46

n ** ** 43

Mean (SD) ********** ********** 1.3 (1.49)

Swollen and tender joints

n ** ** 43

Change from

baseline to final

visit. Mean (SD)

************** ************ 3.9 (22.71)

HAQ pain

N ** ** 43

CONFIDENTIAL




Change from

baseline to final

visit. Mean (SD)

************** ************* 1.37 (30.04)

HAQ disability index

N ** ** 43

Change from

baseline to final

visit. Mean (SD)

************* ************* 0.02 (0.408)

SF36 physical component summary

N ** ** 43

Change from

baseline to final

visit. Mean (SD)

*********** *********** 0.30 (8.97)

4.14 The safety data presented by the manufacturer were from a

pooled analysis of adverse events reported in patients in the

pegloticase 8 mg every 2 weeks groups in the 2 phase III trials

(C0405 and C0406), and long-term safety data from an open-label

extension study (C0407).

4.15 The pooled analysis showed that serious adverse events were

more frequent (24%) in patients in the pegloticase group than in

the placebo group (12%). Similarly, the rate of adverse events

leading to discontinuation was 18% in the pegloticase group and

2% of patients in the placebo group. The most commonly reported

adverse events were gout flare (76% in the pegloticase group and

81% in the placebo group), infusion-related reactions (26% in the

pegloticase group and 5% in the placebo group despite

prophylaxis against infusion-related reactions), headache (9% in

the pegloticase group and 9% in the placebo group), and nausea

(12% in the pegloticase group and 2% in the placebo group).

4.16 All patients who completed C0405 or C0406 were invited to

participate in the open-label extension study C0407 of up to

30 months duration. The primary objective of the open-label

extension study was to assess the long-term safety of

pegloticase. A secondary objective was to evaluate the treatment

CONFIDENTIAL




effects of pegloticase in patients who continued to receive active

treatment from the phase III trials and in those originally

randomised to placebo, and the duration of benefit. Outcomes

included the plasma and serum uric acid response, tophus

response, incidence and frequency of gout flares, swollen joint

count (SJC), tender joint count (TJC), SF-36, HAQ and Clinical

global assessment of disease activity.

4.17 A total of 151 patients entered the open-label extension study

(C0407). Of these patients, 57 had received pegloticase 8 mg

every 2 weeks, 53 had received pegloticase 8 mg every 4 weeks,

and 39 patients had received placebo in the previous trials.

Two patients (3%) who had previously received pegloticase 8 mg

every 4 weeks selected the option of not receiving any further

pegloticase but participated in the study for observation only.

Twenty three out of 39 patients who had received placebo in

previous trials, opted for pegloticase 8 mg every 2 weeks regimen

while remaining 16 received pegloticase 8 mg every 4 weeks in

the open-label extension study (C0407).

4.18 The regimen switches (pegloticase 8 mg every 2 weeks to every

4 weeks and vice versa) were permitted i) after week 25 and

ii) once the results of the double-blind trials were available. The

manufacturer did not provide a summary of why patients switched

and how many switched because of loss of pegloticase efficacy.

4.19 Overall safety observations in the open-label extension were

consistent with those reported in the 2 phase III trials, suggesting

no cumulative risk with continued exposure to pegloticase. Gout

flare was the most commonly reported adverse event, occurring in

71.1% of patients. But the overall percentage of patients with gout

flare decreased throughout the study, particularly in the

CONFIDENTIAL




responders to pegloticase treatment group. Infusion reactions

were the second most commonly reported adverse events,

occurring in 34.6% of patients. Infusion reactions were less

frequent in pegloticase responders. There were 4 deaths in the

open-label extension study, 1 (from sepsis) while receiving

pegloticase, 3 (from cellulitis/sepsis, pneumonia, worsening of

myelodysplastic disorder) at various times after the last dose of

pegloticase.

4.20 A total of 51 (34.2%) patients experienced 106 serious adverse

events across all treatment groups. Of these, 13 were considered

to be possibly or probably related to pegloticase; 11 infusion

reactions, 1 event of nephrolithiasis (kidney stone) and 1 event of

skin necrosis. The remainder were considered unlikely to be

related to the study drug and were consistent with the high degree

of pre-existing comorbidities and polypharmacy that characterised

the study population and the chronic gout refractory to

conventional treatment population more generally.

4.21 A completed non-randomised, non-controlled, open-label,

multicentre re-exposure trial (NCT00675103) (C0409) was

identified by the ERG that was not included in the manufacturers

submission. This small-scale trial evaluated efficacy and safety

outcomes in patients receiving a 24-week course of pegloticase

whose last exposure to pegloticase was at least 1 year before trial

entry. The manufacturer provided a brief synopsis of this trial,

which evaluated efficacy and safety outcomes in a small number

(n=7) of patients who were re-exposed to a subsequent course of

pegloticase treatment after an initial course of treatment.

***************************************************************************

***************************************************************************

***************************************************************************

CONFIDENTIAL




***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

***************************************************************************

**.

4.22 Pre-planned subgroup analyses of the individual phase III trials

(C0405 and C0406) and of the pooled data for treatment

responder were undertaken by the manufacturer according to sex,

age (55 years or younger, older than 55 years), body mass index

(30 kg/m2 or less, more than 30 kg/m2), absence or presence of

tophi, disease duration (less than 5 years, 5 years or more), and

baseline HAQ disability index score (1 or less, more than than 1)

creatinine clearance (less than 50 ml/min, 50 ml/min or more) and

antibody status. These compared plasma uric acid responders in

the pegloticase and placebo groups using Fishers exact test (see

table 6.12 in the manufacturers submission). The manufacturer

stated that results for the pooled data did not indicate any clear

pattern or trend of improved efficacy in any subgroup.

4.23 The ERG was satisfied that all available phase III trials were

included in the submission but noted that in the two phase III trials

a higher proportion of patients in the pegloticase every 2 weeks

arms had dropped out at the end of the study (30% in C0405 and

28% in C0406) than in the placebo arms (5% in C0405 and 13%

in C0406). The ERG stated that the manufacturer did not explain

CONFIDENTIAL




or adjust for the imbalances in these the drop-out rates,

particularly in the patients who dropped out because of adverse

events in both trials.

4.24 The ERG highlighted the manufacturers approach of pooling data

from the two phase III trials and pointed out that simple pooling of

data may yield counterintuitive or false results and that a meta-

analysis would have been a better approach to combining the

results of the 2 trials. The ERG performed a fixed and random

effects exploratory meta-analysis for plasma uric acid responder

status and tophi resolution. These outcomes were selected

because plasma uric acid responder status was the primary

efficacy outcome and tophi resolution was a key driver in the cost-

effectiveness model. The ERG stated that the software used to

perform exploratory meta-analysis automatically applied a

correction of 0.5 to the placebo arms of both trials if no events

were observed and noted that this may introduce uncertainty to

the results.

4.25 The ERGs meta-analysis for the primary efficacy endpoint of

plasma uric acid responder status confirmed that response was

significantly greater in the pegloticase 8 mg every 2 weeks

treatment group than the placebo group (relative risk [RR] 18.99,

95% confidence interval [CI] 2.69 to 133.94, p=0.003). When a

fixed effects model was applied, the relative risk was very similar

(RR 19.01 95% CI 2.69 to 134.24). It was not possible to calculate

a precise relative risk in the pooled data because there were no

events in the placebo arm. However, using a similar correction to

that applied in the meta-analysis (see section 4.24) the relative

risk was 37.35 (95% CI 2.35 to 549.24).

CONFIDENTIAL




4.26 At the clarification stage the manufacturer stated that tophi

resolution responses were conducted on the tophus-evaluable

population, which included patients with a tophus at baseline and

patients who developed new tophi during the trial. This means

that the number of tophus-evaluable patients was greater than the

number of patients with tophi at baseline. The ERG considered

that it was unclear how data relating to the total size of the

tophus-evaluable population at final visit had been derived and

why the number of patients with tophi at baseline had not been

used in the analyses of tophi resolution at final visit. So the ERG

conducted 2 separate exploratory meta-analyses for complete

tophus resolution, based on the tophus-evaluable population and

the number of patients with baseline tophi.

4.27 The results of the ERGs meta-analysis of tophus resolution

based on a tophus-evaluable population showed a relative risk of

4.17 (95% CI 1.25 to 13.92), which statistically significantly

favoured pegloticase over placebo. For comparison, the ERGs

calculation of the relative risk based on the simple pooled data

was 5.45.The ERGs exploratory meta-analysis based on number

of patients with baseline tophi generated a relative risk of 3.62

(95% CI 1.07 to 12.27). With the fixed effects model the combined

relative risk was broadly similar (RR 4.04, 95% CI 1.19 to 13.71).

The relative risk calculated for the simple pooled data was 4.91.

The ERG commented that basing the calculation on the baseline

number of patients with tophi better reflected the intention-to-treat

population than using the manufacturers preferred tophus-

evaluable population.

4.28 The ERG also highlighted that the impact of repeated courses of

pegloticase on uric acid levels, secondary outcomes,

immunogenicity and adverse events was not clear from the

CONFIDENTIAL




manufacturers submission. It stated that limited evidence from

the small scale re-exposure trial (C0409) suggested that there

***************************************************************************

****************************************************************. The

ERG also stated that it was not clear whether the benefits of

pegloticase would be maintained after pegloticase treatment

stopped or whether maintenance treatment with other urate-

lowering drugs would be successful in maintaining low uric acid

levels and other benefits in the long-term.

5 Comments from other consultees

5.1 Professional organisations pointed out that gout is poorly

understood and managed in primary care. GPs often fail to

properly titrate the first course of urate-lowering treatment

because of an initial increase in flare rates and incorrectly classify

patients as intolerant. In particular, allopurinol intolerance is often

reported when it has simply been associated with triggering a flare

of gout, without any attempt to perform desensitisation.

Professional and patient organisation emphasised that

desensitisation should always be tried first in patients before

categorising them as allopurinol intolerant. Uricosuric drugs like

probenecid, sulphinpyrazone and benzbromarone (used outside

their marketing authorisation but recommended by EULAR) are

rarely used in clinical practice in the UK and are considered of

limited value. Professional organisation also highlighted the

uncertainty in the evidence base about the best sequence of

treatment after allopurinol. Professional organisations stated that

adjunctive treatment with oral, intramuscular or intra-articular

corticosteroids is often overlooked in the clinical practice.

CONFIDENTIAL




5.2 Professional organisations stated that that the treatment of gout in

patient with significant comorbidities, for example severe kidney

disease, cardiac failure, and haematological disorders is

particularly challenging because many of the available treatments

are contraindicated in these patients. Professional organisations,

while acknowledging the lack of data, hope that pegloticase may

be particularly useful in this group of patients because of its

unique mode of action compared with existing therapies. However

they highlighted that the trials did not report on this subgroup of

patients with significant comorbidities, despite including patients

with moderate renal impairment and cardiovascular comorbidities.

5.3 Professional organisations were also concerned about the

external validity of the trial data given that the trials were relatively

short but gout is a lifelong condition. There was a particular

concern because of a potential immune response to pegloticase,

which may reduce its efficacy and cause adverse reactions.

5.4 Patients group said preventing acute episodes of gout was the

most important treatment outcome and that reducing severity of

pain, swelling, tenderness, joint erosion and the size and number

of tophi were also important treatment outcomes. In particular,

they said that reducing the severity of gout attacks may mean that

patients can conduct day-to-day activities and improve their

quality of life. But members of the patient groups had no personal

experience of pegloticase because the trials were outside the UK.

5.5 A potential disadvantage of pegloticase identified by the patient

groups was frequent intravenous infusions, which take between 2

and 4 hours each time. Patients would have to weigh up how

often attacks happen against the benefit of treatment when

deciding whether to be treated. The patient groups also

CONFIDENTIAL




highlighted concerns about the reported incidence of infusion

reactions (2637.5%) which led to up to 10% of patients stopping

pegloticase, and delayed hypersensitivity reactions despite

corticosteroid or antihistamine prophylaxis. The patient groups

further highlighted a high incidence of gout flare, lack of clarity

about the optimal duration of treatment and the development of

anti-pegloticase antibodies, leading to loss of efficacy. Additionally

they said there is a potential risk of cardiovascular adverse

effects, particularly in a population at increased risk of

cardiovascular disease because of the association between gout

and the metabolic syndrome.

5.6 The patient groups said they see pegloticase as a bridging option

in patients with refractory tophaceous gout, for the first 6 months

of treatment, to allow rapid resolution of tophi. Patients could then

revert to standard treatment. The patient groups also emphasised

on the need of data collection to determine long term tolerability

and toxicity.

6 Cost-effectiveness evidence

6.1 The manufacturer identified 1 economic evaluation of pegloticase

in patients with refractory chronic gout performed from a US

healthcare perspective (Wang et al. 2012) but did not consider it

relevant for this appraisal. The manufacturer therefore submitted

a de novo model comparing pegloticase with best supportive care

in patients with chronic gout refractory to xanthine oxidase

inhibitors. Each branch of the decision tree structure (see figure 1)

is coupled with a Markov model for extrapolation to the 20-year

base-case time horizon. There are 4 health states in the Markov

model based on patients serum uric acid levels:

below 360 micromol/litre

CONFIDENTIAL




360 or higher but below 480 micromol/litre

480 or higher but below 600 micromol/litre

600 micromol/litre or higher.

Death is also included as a Markov state in the model.

Figure 1 Decision analytical structure

6.2 In the pegloticase arm, patients are separated into responder,

non-responder and non-completer groups:

Responders are defined in the manufacturers submission as

people responding to pegloticase, that is, maintaining serum uric

acid levels below 360 micromol/litre. These patients are

assumed to gain the most clinical benefit from pegloticase. After

the treatment course with pegloticase (lasting a maximum of

6 months in the base-case analysis), patients are switched to

maintenance treatment with either allopurinol or febuxostat.

Based on the trial data of persistence with pegloticase (68%)

and response rate (60%), the manufacturer estimated that

40.8% of the cohort would be classified as responders.

CONFIDENTIAL




The non-responders are defined as those who discontinue

pegloticase treatment because serum uric acid levels exceed the

limit set out in the summary of product characteristics: increase

to above 360 micromole/litre on 2 consecutive observations. In

the base-case analysis, patients were assumed to be identified

as non-responders after 2 months of pegloticase treatment.

These patients, estimated to be 27.2% of original cohort, were

switched to best supportive care. Based on the clinical trial data,

it is assumed that some clinical benefit was still obtained in this

patient group.

The non-completer group was defined in the manufacturers

submission as patients who are non-persistent to pegloticase

treatment, but subsequently the manufacturer clarified that this

included patients who failed to complete the full 6 month

treatment course for any reasons, primarily due to adverse

events, but could also include other reasons such as patient

dislike of intravenous administration. These patients are

assumed to not gain any clinical benefit from pegloticase

treatment although pegloticase treatment costs are incurred. In

the base-case analysis, it was assumed that a non-completer

could be identified by 1 month of pegloticase treatment. In the

manufacturers model 32.0% of patients do not complete

pegloticase treatment and are switched to best supportive care

after 1 month.

6.3 In the manufacturers model patients in the best supportive care

arm are not treated with pegloticase. This group was modelled

from the placebo treatment group in the clinical trials.

6.4 The model assumed that all patients who respond to pegloticase

treatment will progress to maintenance treatment with either

allopurinol (70%) or febuxostat (30%). Discontinuation of

CONFIDENTIAL




maintenance treatment was modelled with data derived from an

observational study of allopurinol (Annemans et al. 2008) and

from a randomised trial of febuxostat (EXCEL). Patients who

discontinued maintenance treatment (non-persistent patients)

were switched to best supportive care. It was assumed that

maintenance treatment was effective in maintaining the serum

uric acid levels achieved by pegloticase treatment, and that with

non-persistence with maintenance treatment serum uric acid

levels returned to baseline levels. The probability of becoming

non-persistent on maintenance treatment, and therefore

progressing to best supportive care, was modelled using a

constant hazard for febuxostat, and a decreasing hazard for

allopurinol. The rate of discontinuation for febuxostat was derived

by taking the average proportion (434/590=26%) who were non-

persistent at the end of the 2-year EXCEL trial across the 2 doses

(80 mg/day and 120 mg/day) and assuming a constant hazard

(exponential survival curve) to convert this proportion to a monthly

rate of 1.3%, which is equivalent to an annual rate of 14%. The

rate of discontinuation for allopurinol was derived by fitting a

Weibull survival curve to retrospective 2-year observational data

from 7443 UK patients with chronic gout, of whom 89% were

receiving allopurinol. In the economic model, the manufacturer

stated that a Weibull curve was fitted to the second year of data

from the trial to reflect long-term persistence (as opposed to

persistence from the start of treatment).

6.5 The baseline characteristics of the modelled population were

defined according to age (56 years), serum uric acid level

(9.6 mg/dl) and baseline utility values (0.6) based on the baseline

demographic characteristics of the 2 phase III trials. No natural

disease progression was modelled because tophaceous gout was

considered the final stage of gout.

CONFIDENTIAL




6.6 Each of the responder, non-responder and non-completer groups

has a distinct set of health states with no transitions allowed

between the groups. For each of these groups, the model has

health states defined according to the treatment being given and

the serum uric acid level at that time point. In the comparator arm,

all patients receive best supportive care for the duration of the

model so the health states are defined solely according to serum

uric acid level.

6.7 The model estimated the distribution of patients across the 4

serum uric acid levels by assuming a normal distribution around

the mean serum uric acid level for each group in the pegloticase

arm (responders, non-responders, non-completers) and for the

comparator arm population as a whole. Transitions to the death

state are possible at any time and do not depend on the treatment

being given or the patients serum uric acid level and are therefore

the same across the whole population of the model.

6.8 Treatment effectiveness is captured within the model by changes

in 4 patient-related outcomes: serum uric acid, quality of life,

frequency of flares and tophi resolution. Patient-level data were

available from the 2 replicate trials (C0405 and C0406) for all

4 outcomes. Mean values for serum uric acid, frequency of flares,

tophi resolution and quality of life are presented in tables 7.2, 7.3

7.4 and 7.7 of the manufacturers submission for the pegloticase

responder, pegloticase non-responder and best supportive care

groups. No outcomes are presented for the pegloticase non-

completer group.

6.9 In the model, serum uric acid levels in responders were based on

data from the pegloticase arm for the first 6 months and then

maintained at 0.17 mg/dl during maintenance treatment. After

CONFIDENTIAL




patients became non-persistent with maintenance treatment,

serum uric acid levels were based on trial data from the

pegloticase non-responder group for months 6 to 12 of the model

and on the baseline level for the whole trial cohort for the

remainder of the model. Serum uric acid levels in non-responders

were based on data from non-responders in the pegloticase arm

of the trials for 2 months and then on data from the placebo arms

of the 2 phase III trials for the remaining 4 months once the

patients have progressed to best supportive care. Serum uric acid

levels in non-completers were based on data from the placebo

arms of the 2 phase III trials for the full 6 months. The baseline

serum uric acid level was modelled beyond 6 months for non-

responder and non-completer patients.

6.10 Frequency of flares in the pegloticase arm was based directly on

trial outcomes in the first 6 months for responders receiving

pegloticase. Frequency of flares in the pegloticase non-responder

group was based on trial data from non-responders for 2 months

and then estimated on the serum uric acid level once the non-

responders progress to best supportive care. Frequency of flares

in the pegloticase non-completer group was based on trial data

from the placebo arm for 1 month and then on the serum uric acid

level once the non-completer progress to best supportive care.

Frequency of flares for patients receiving best supportive care in

the comparator arm in the first 6 months was derived directly from

the trial data and later on based on the serum uric acid level.

6.11 Tophi resolution was assumed to increase linearly over the first

6 months to the level seen at month 6 in the pooled data of the

2 key trials. The resolution of tophi achieved during pegloticase

treatment was assumed to be maintained in 100% of responders

and 50% of non-responders for the entire time-horizon of the

CONFIDENTIAL




model. For non-completers and patients in the best supportive

care arm of the model tophi resolution achieved at 6 months in the

pooled analysis of the placebo arms of the trials was modelled

and maintained for entire time horizon of the model.

6.12 Short term utility estimates (first 6 months) were based on SF-6D

utilities derived from the SF-36 data collected in the pegloticase

phase III trials and validated against the EQ-5D. Modelled utility

values were corrected for differences in baseline utility of different

patient groups. No disutilities associated with adverse drug

reaction were included in the model. The manufacturer said this

was because utility data for the entire duration of pegloticase

treatment in the model were collected in the trial, and therefore

the disabilities associated with adverse effects would already be

captured.

6.13 Long-term utilities were based on estimates derived from

guidance on febuxostat for hyperuricaemia in gout (NICE

technology appraisal guidance 164) and Scottish Medicines

Consortium guidance (367/10) on febuxostat. These values were

based on a European observational study in chronic gout patients

with health-related quality of life and utility measured by EQ-5D

and utilities for each serum uric acid state derived by regression

analysis. The utility values applied in the model after 6 months

(and in some patients before 6 months, for instance in non-

responders after 2 months and in non-completers after 1 month)

were based on a combination of serum uric acid level, frequency

of flares and tophi resolution.

6.14 The anticipated acquisition cost of pegloticase was used in the

model. Other costs included the costs of intravenous drug

administration, prophylaxis for acute gout flares and infusion

CONFIDENTIAL




reactions, adverse event costs, refractory chronic gout

management, treatment for acute flares, and tophus surgery.

Resource use for the management of severe debilitating

refractory chronic gout according to serum uric acid level was

estimated by clinical expert survey and the in-depth clinician

interview. Drug cost incurred for prophylaxis for acute gout flares

and infusion reactions, adverse event costs, refractory chronic

gout management, acute flares treatment were based on costs in

the British National Formulary (BNF) 63 All other resource use

cost was estimated using NHS reference costs and Health

Related Groups for 201011. The resource use for 2 adverse

drug effects (infusion reactions and vomiting) was modelled for

the pegloticase arm and assumed 0 for best supportive care,

which was assumed to consist of standard medical care with

NSAIDS, colchicine and corticosteroids but no urate-lowering

treatment. However, a cost for drug treatment associated with the

best supportive care comparator was not included in the

economic model in the original submission. The cost used in the

model of managing an episode of acute flare was 295 based on

a Scottish Medicines Consortium guidance (637/10) on

febuxostat.

6.15 In the manufacturers submission the base-case ICER for

pegloticase compared with best supportive care was reported to

be 29,946 per QALY gained based on incremental costs of

9466 and an estimated QALY gain of 0.316. The ICER

generated by probabilistic analysis was 29,833 (after correction

by the ERG for an error identified in the computer program used

to run the probabilistic sensitivity analysis).

6.16 The deterministic sensitivity analyses conducted by the

manufacturer showed that the cost effectiveness of pegloticase

CONFIDENTIAL




was particularly sensitive to changes in baseline levels of serum

uric acid, the disutility associated with higher serum uric acid

levels and patients age. The sensitivity analyses also show that

the disutility associated with higher serum uric acid levels is a

significant driver of cost effectiveness, which is important given

the significant uncertainty about this assumption. The cost-

effectiveness results were also fairly sensitive to changes in the

utility value for patients with serum uric acid level under the target

value and the baseline utility value. The cost-effectiveness results

were also moderately sensitive to the parameter values for

treatment efficacy and persistence with pegloticase treatment.

6.17 The scenario analyses showed the sensitivity of the ICER to

several structural changes in the models clinical pathways. The

most sensitive variables were related to how long patients receive

pegloticase for and how soon non-responders stop treatment with

pegloticase. ICERs in the scenario analyses for pegloticase

compared with best supportive care ranged from 19,817 per

QALY gained (assuming duration of treatment with pegloticase

5 months instead of 6 months as in the base case) to 38,025 per

QALY gained (assuming evaluation of non-completer after

3 months instead of 1 month as in the base case).

6.18 At the clarification stage, several assumptions in the

manufacturers model were questioned. The model assumed that

all patients who responded to pegloticase treatment will progress

to maintenance treatment with either allopurinol (70%) or

febuxostat (30%). However, the ERG stated that this treatment

sequence would not have been appropriate for patients in whom

conventional urate-lowering therapies are contraindicated or not

tolerated. Furthermore the ERG was concerned that the original

model did not include any drug costs for best supportive care. The

CONFIDENTIAL




manufacturer responded with a revised model in which 10% of the

responders were assumed to switch to best supportive care,

instead of maintenance treatment with xanthine oxidase inhibitors,

after pegloticase treatment. The revised model also included drug

costs for best supportive care. Best supportive care treatment was

considered to consist of an NSAID (for instance naproxen 500 mg

per day) in about 90% of patients, colchicine (for instance

1000 mg per day) in about 10% of patients and corticosteroids (for

instance prednisolone 1015 mg per day) in about 75% of

patients. The deterministic results for this revised base-case

analysis (referred to as the revised model) showed an incremental

cost of 9452 and incremental QALY of 0.305, compared with

best supportive care resulting in an ICER of 31,027 per QALY

gained. A probabilistic sensitivity analysis (after correcting for an

error identified in the computer program used to run the

probabilistic sensitivity analysis) of 10,000 samples using the

revised assumptions yielded a mean ICER of 31,031 per QALY

gained (incremental cost of 9491 and a QALY gain of 0.306).

6.19 In response to additional concerns by the ERG, the manufacturer

also provided a number of additional deterministic sensitivity

analyses, including:

exploring the effect of using alternative data to calculate

persistence with allopurinol

incorporating costs of G6PD screening

more frequent serum uric acid level monitoring

pharmacy time with a specialist consultation for pegloticase non-

responders and non-completers at the start and end of the

treatment

assuming no extra decrease in utility in patients with tophi,

based on the assumption that the impact of tophi on quality of

CONFIDENTIAL




life had already been captured in the 4 serum uric acid health

states

assuming no correlation between serum uric acid levels and

utility

different mortality rates.

The revised ICERs were particularly sensitive to assumptions that

tophi resolution influences utility and that serum uric acid levels are

correlated with utility. The ICER increased to 43,614 per QALY

gained when no extra decrease in utility was assumed in patients

with tophi, and to 38,535 when serum uric acid levels were

assumed not to be correlated with utility.

6.20 The ERG questioned several assumptions in the manufacturers

submission. It noted that in the 2 key trials, it took up to ******** for

all patients who eventually lost response or withdrew from

pegloticase to be identified, however the model assumed that

responders can be identified after 2 months of pegloticase

treatment. Additionally, the ERG stated that the group termed

non-completers by the manufacture would be difficult to identify

in clinical practice. It also noted the assumption that non-

completers could be identified in the first month of treatment was

arbitrary because there were no data on time to stopping

treatment from the trials.

6.21 The ERG suggested that an alternative modelling approach could

have classified each non-completer patient as a non-responder

and incorporated their outcomes within the mean result for non-

responders. This would have been a more accurate reflection of

the trials.

6.22 The ERG noted that in the revised model the proportion of

pegloticase responders who progressed to best supportive care

CONFIDENTIAL




without having intervening maintenance treatment was assumed

to be 10% but the proportion of trial patients who were eligible to

participate because of intolerance or contraindication to

allopurinol was ***** in trial C0406 and ***** in trial C0405.

6.23 The ERG commented that the revised model assumed that

patients switching to best supportive care have a rapid return of

high serum uric acid levels and the only treatment benefits

maintained are those associated with tophi resolution. This may

have underestimated treatment benefits in the responders who

were intolerant to allopurinol or for whom allopurinol is

contraindicated. It said that the rise in serum uric acid could be

assumed to be more gradual, but that this remained uncertain

because there were no data.

6.24 The ERG was concerned that that the length of time on treatment

in clinical practice could be greater than the 6 months assumed in

the model. It noted that the draft summary of product

characteristics did not limit treatment to 6 months but also

acknowledged the lack of data to support long-term treatment.

The combined data from the 2 key trials (C0405 and C0406) and

the open-label extension study (C0407) showed that the mean

number of treatments received by responders was ****

(median=**, SD=*****). A Kaplan-Meier plot of the time to stopping

treatment for responders to pegloticase all 3 trials showed that the

proportion continuing to receive pegloticase treatment was fairly

high at over *** until around *******, with

*****************************************. The ERG also noted that in

the open-label extension study patients were allowed to continue

pegloticase treatment for up to a maximum of 30 months

(2.5 years).

CONFIDENTIAL




6.25 The economic model assumes that the benefits of pegloticase

treatment in responders during the 6-month treatment period can

be maintained in the long term by treatment with allopurinol or

febuxostat. The ERG considered that this was uncertain because

there was no direct evidence on persistence with either allopurinol

or febuxostat in the population receiving maintenance treatment in

the economic model (patients with severe debilitating chronic

tophaceous gout who had not responded to or are intolerant to

xanthine oxidase inhibitors and who had responded to treatment

with pegloticase). The ERG noted that in the manufacturers

original model 34% of the incremental QALY gains were accrued

more than 10 years after starting pegloticase treatment. The ERG

considered that the extrapolation of benefits over such a long

period was a significant area of decision uncertainty because no

direct evidence had been presented by the manufacturer to show

that the serum uric acid levels after response to pegloticase

treatment can be maintained by treatment with allopurinol or

febuxostat treatment in the long term.

6.26 The ERG also had concerns about the survival model used to

extrapolate persistence with allopurinol, and noted that in the

scenario analysis the ICER was sensitive to the rate of

discontinuation of allopurinol. The ERG noted that an alternative

approach to modelling allopurinol persistence could be a Weibull

fit for all the 2-year data available from the UK observational study

(Annemans et al. 2008) as opposed to the Weibull fit for only the

second year data, as in the original model. The ERG noted that a

sensitivity analysis conducted by the manufacturer at the

clarification stage using this approach increased the ICER to

37,981 per QALY gained from the revised base-case ICER of

31,027 per QALY gained.

CONFIDENTIAL




6.27 The ERG noted that published data from modified intention-to-

treat analysis of the trials (see section 4.5) were not used to

calculate the clinical effectiveness in the model, and that there

was additional analysis of the trial data to inform the health

economic model. Primary end points in the trials were based on

plasma uric acid level while in the economic model the end points

were based on serum uric acid. However, the ERG noted that

plasma and serum uric acid levels had a degree of correlation and

that serum uric acid levels would likely be used in clinical practice.

The ERG agreed that long-term maintenance of serum uric acid

below 360 micromol/litre could be expected to result in clinically

meaningful changes in patient-related outcomes, although it noted

that the BSR guideline recommended maintaining serum uric acid

levels below 300 micromol/litre.

6.28 The ERG noted that the effectiveness of tophi resolution (in terms

of proportion of patients with resolved tophi) was applied to all

patients in the model and not only to the subgroup with tophi at

baseline. The ERG considered that the model might have

overestimated the treatment benefit of tophi resolution and

therefore biased the analysis in favour of the pegloticase arm. The

ERG also commented that the proportion of patients receiving

pegloticase in clinical practice who have tophi at baseline may be

higher than what was observed in the trials because of the

wording of the draft license indication, which includes the term

tophaceous gout.

6.29 The ERG had concerns that the method for calculating utility from

a combination of serum uric acid level, frequency of flares and

tophi resolution may result in quality of life improvements being

double counted because all 3 are likely to be correlated in an

individual. For example, tophi resolution is likely to be correlated

CONFIDENTIAL




with low serum uric acid levels, but the model assumes that the

probability of having tophi resolution is the same irrespective of

the serum uric acid level, and applies a utility benefit for both tophi

resolution and low serum uric acid levels.

6.30 The ERG stated there was a similar risk of double counting when

capturing both the benefits of reduced flares and lower serum uric

acid level, because the frequency of flares is assumed to be

related to the serum uric acid level but in the model both are

treated as independent factors in determining utility. This type of

double counting was more problematic because both utility gains

were applied to all patients from 6 months and to patients in the

pegloticase arm of the model who receive best supportive care

before 6 months. The ERG suggested the potential for double

counting benefits in this manner may have been avoided if the

data on the relationship between serum uric acid and utility had

been properly adjusted to take into account the frequency of flares

as a confounding factor, providing an estimate of the disutility per

serum uric acid state in the absence of flares.

6.31 The ERG was concerned about the assumption that the resource

use would be greater in patients with a higher serum uric acid

level over and above the difference determined

Date post:	08-Jan-2016
Category:	Documents
Upload:	suryansan
View:	11 times
Download:	0 times

62574

Documents