- Email: [email protected]
Changes in the Prescribed Daily Doses of Transdermal Fentanyl and Transdermal Buprenorphine During Treatment of Patients with Cancer and Noncancer Pain in Germany: Results of a Retrospective Cohort Study
.... i
•~ *
'~ ~ Ciinicai ~Therapeutics/Vo|urne27, Number ! 7, 20Q$ ~ :~ :
;~ i~:::
:~:i
' '~~
Changes in the Prescribed Daily Doses of Transdermal Fentanyl and Transdermal Buprenorphine During Treatment of Patients with Cancer and Noncancer Pain in Germany: Results of a Retrospective Cohort Study Reinhard Sittl, MD1; Marl< Nuijten, MD, PhD2; and Barbara Poulsen Nautrup, PhD 3
1Pain Clinic, Universityof Erlangen, Erlangen, Germany;2MEDTAPInternational, Amsterdam, the Netherlands; and 3Gruenentkal GmbH, Aachen, Germany ABSTRACT Background: During long-term therapy with strong opioids (step III opioids according to the World Health Organization [WHO] analgesic ladder), dose increases are often necessary because of deterioration of the primary disease or development of tolerance. Objective: The purpose of this study was to compare changes in dosages of transdermal (TD) fentanyl and TD buprenorphine in patients with cancer and noncancer pain. Methods: In a retrospective study, patients with cancer and noncancer pain being treated with TD fentanyl or TD buprenorphine for at least 3 months between January 2001 and December 2003 were identified from the IMS Disease Analyzer-mediplus database, which contains all patient-related data documented from 400 medical practices in Germany. The indications for treatment were define, t according to the International Classification of Diseases, l Oth Revision, and included neoplasm (cancer groups), and osteoarthritis, low back pain, and osteoporosis (noncancer groups). The cohort patients were considered to have comparable pain intensity because they had received similar analgesic premedication classified according to steps I to III of the WHO analgesic ladder (cohort groups). The mean prescribed daily doses on first and last prescription were documented, and the mean percentile increases were calculated over the whole treatment duration and per day. Additionally, the mean percentile intraindividual increases (on a per-patient basis) were estimated. Results: The cohort groups consisted of 448 patients with noncancer pain and 446 patients with cancer pain (552 women and 342 men; mean age, 74 years; range, 25-101 years). The mean percentile increases in dosages over the whole treatment duration and adjusted per day were significantly higher in patients
1022
., ,~,
taking TD fentanyl (P < 0.05). Differences were even greater for the mean percentile intraindividual increases per day, which totaled 0.42% and 0.17% for cancer patients taking TD fentanyl and TD buprenorphine, respectively; corresponding values were 0.25% and 0.09% in noncancer patients (P < 0.001). Conclusions: This retrospective analysis showed a significantly higher increase in the mean daily doses of TD fentanyl as compared with TD buprenorphine. The results must be verified in prospective, randomized clinical studies. (Clin Ther. 2005;27:1022-1031) Copyright © 2005 Excerpta Medica, Inc.) Key words: analgesic, buprenorphine, tolerance, fentanyl, transdermal.
INTRODUCTION Opioids are the mainstay of treatment for chronic moderate to severe pain. 1,2 The goal of pain management with opioids is to provide optimal analgesia with a minimum of adverse effects. The World Health Organization (WHO) guidelines on cancer pain management stipulate adapting the dosage of analgesics to the degree of pain. 3 Accordingly, there are no standard doses for opioid drugs. The "right" dose is the dose that relieves the patient's pain. 3 This work was presented in part as an abstract at the International Society for Pharmacoeconomics and Outcomes Research 10th Annual International Meeting, May 15-18, 2005, Washington, DC.
Accepted fiorpublication June 2, 2005. Express Track online publication June 28, 2005. doi:l 0.1016/j.clinthera.2005.06.024 0149-2918/05/$19.00 Printed in the USA. Reproduction in whole or part is not permitted. Copyright © 2005 Excerpta Medica, Inc.
~
~
. .
.
.
.
.
Volume-27 Number 7 .
During long-term therapy with strong opioids, dose increases are often necessary because of deterioration of the primary disease, development of tolerance, or both. 4 Analgesic tolerance is defined pharmacologically as a reduced analgesic effect of opioids after repeated administration or the need for higher doses to maintain the same effect. 5 Tolerance develops over hours, days, or weeks of opioid application. 6 Opioid tolerance mainly develops on the level of the target cells; 2 well-described mechanisms include opioid receptor uncoupling from G-proteins and receptor downregulation. 7 Clinically, using high dosages of opioids to overcome tolerance may result in a greater risk of adverse effects because of an increase in toxic opioid metabolites, s Tolerance to the opioid's effects (therapeuuc and toxic) develops at different rates. Among the adverse effects, the development of tolerance is mimmal for constipation and miosis, whereas tolerance occurs rapidly to nausea, vomiting, sedation, euphoria, and respiratory depression, s For chronic pain treatment, the transdermal (TD) opioid formulations of fentanyl* and buprenorphine~ have been used in many countries. Fentanyl is a full agomst at p-opioid receptors and was introduced into medical practice in the 1960s. It is a lipid-soluble synthetic opioid that can be delivered in a TD controlledrelease formulation for up to 72 hours. 9 TD fentanyl has proved effective in the treatment of cancer and noncancer pain. ~°-~3 TD buprenorphine is a centrally acting analgesic that was introduced to European markets >20 years ago and is some 25 to 50 times more potent than morphine on a weight-by-weight basis34,~s The compound has a high affinity for opioid receptors, acting as a partial agonist at la-opioid receptors and as an antagonist at ~:-opioid receptors. Beginning in 2001, the TD buprenorphine formulation has been launched in many European countries. The efficacy and safety of TD buprenorphine have been demonstrated in randomized, placebo-controlled studies in cancer and noncancer patients. 16-~8 TD formulations are new treatment options for longterm pain management in patients who require stable dosing. The advantages of TD therapeutic systems include maintenance of constant plasma levels, which have a favorable impact on the incidence of adverse ef~Trademark: Durogesic® (Janssen Pharmaceutical Products kP, Neuss, Germany). 1"Trademark: Transtec® (Gr0nenthal GmbH, Aachen, Germany).
fects. Some evidence indicates that TD opioids may improve patients' compliance and quality of life during long-term therapy39,2° For TD fentanyl, cancer pain studies have documented significant increases in the dosages during treatment. Mean initial versus mean final dosages were 38.7 versus 239.6 pg/h (mean treatment duration, 1095 days), ~° respectively, and 44.0 versus 178.0 lag& (mean treatment duration, 90 days), l~ In studies of noncancer pain, dosage increases were also reported from 25.8 to 67.3 lag& (mean treatment duration, 304 days) 12 and from 48.0 to 90.0 tlg/h (treatment duration, 365 days). 13 However, these studies did not differentiate the reasons for increased dosages, whether dose titration, deterioration of the underlying disease, or tolerance development. With buprenorphine treatment, tolerance development has been reported to be low, but these data were not based on the TD bnprenorphine formulation, and no figures on tolerance development were reported34,2~ Literature searches in MEDLINE and EMBASE for the years 1995 through 2005 did not find any study on TD buprenorphine that documented the initial and final dosages (ie, changes in dosages during treatment). Accordingly, we did not find a direct comparison in the literature of dosage changes between TD fentanyl and TD buprenorphine in comparable patients according to pain and underlying disease. The aim o f this retrospective study was to evaluate the prescribed dosages of TD fentanyl and TD buprenorphine that were documented over a period of 3 years in Germany in patients with cancer and noncancer pain (due to osteoarthritis, low back pain, and osteoporosis) and to compare the dosage changes over the course of treatment. PATIENTS A N D METHODS Patient Data
Patient data analyzed in this study were derived retrospectively from the IMS Disease Analyzer-mediplus database, which has been used to investigate disease courses and therapy patterns. 22,23The database contains anonymous information on patients treated in 400 practices of general practitioners and internists in Germany. Patients'being treated with TD fentanyl or TD buprenorphine were analyzed for a documentation period of 3 years, between January 2001 and December 2003. Germany does not require institutional review board waiver of consent for retrospective studies of anonymous data.
Pain treatment was evaluated separately for noncancer and cancer patients. A similar methodology was used as described in a previous study, which analyzed the same database over a similar period of 3 years; however, the previous study covered a different time frame and had a different objective. 23 The indications for treatment were defined according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-1 O) diagnostic codes.24 The following indications were included: for the cancer groups, patients with the ICD-IO codes COO through C97 (neoplasm), and for the noncancer groups, patients with the ICD-10 codes M15 through M19 (osteoarthritis), M40 through M54 (low back pain), and M80 through M85 (osteoporosis). For a direct comparison of the dosages and dosage changes, it was essential to perform the analyses on comparable patients according to pain intensity. Because the analgesic dosage depends on the intensity of pain and is administered and titrated freely within the recommended dose range of the opioid, a comparison was rational only in patients with similar pain intensity. Because-pain intensity was not documented in the database used, patients were assumed to have comparable pain intensity if they had received similar analgesic medication before the patch treatment, as described in detail in a previous study.23 According to the methodology in this study, we defined a subgroup of patients on long-term TD fentanyl or TD buprenorphine treatment as _>3 months, which corresponds to >-30 patch prescriptions, assuming a wearing time of 3 days per patch as recommended. For evaluation of previously received analgesic medication (premedication), all drugs from the European Pharmaceutical Market Research Association (EPhMRA) classifications 2s of N02A (opioids), N02B (other analgesics), and M01A (anti-inflammatory and antirheumatic products) were considered and classified according to step I, II, or III of the WHO ladder for pain treatment.3 The WHO ladder specifies that the first step is a nonopioid, the second is a weak opioid for mild to moderate pain, and the third is a strong opioid for moderate to severe pain. If no analgesics were prescribed previously, this was classified as no premedication. Patients receiving premedication analgesics from different steps of the WHO ladder, such as steps I and III, were classified as belonging to the highest step (eg, step III in the given example).
1024::
Assessmentof Dosages The daily doses were determined directly from the prescribed patch strength. However, when calculating the prescribed daily doses, we considered that because of the label recommendations, patients could apply > 1 patch at a time,26,27 a fact that has often been reported for TD fentanyl.4,1°-12,28,29 However, the database used did not explicitly mention when 2 patches had to be applied concomitantly. Regarding concomitant prescriptions of different patch strengths, double patch applications were taken into consideration only if they were obvious. For example, a concomitant prescription of 100 + 50 lag/h of TD fentanyl was only considered to be a double application if the resulting dosage seemed rational with regard to the previously and subsequently prescribed doses. If a concomitant prescription on 1 day was obviously intended for dose titration, the double prescription was not counted as a concomitant application. The possibility of identifying patients who applied >1 patch was limited for daily doses >175 lag& for TD fentanyl (eg, 2 × 100 pg/h) or >122.5 lag& for TD buprenorphine. It was not possible to identify patients who received prescriptions of ] patch strength but applied _>2 patches (eg, 2 × 50 pg/h).
Changes in Dosages In all patients in the cancer and noncancer cohorts, the dosages prescribed at first and last prescription (regardless of the time until last prescription), as well as the dosages on days 100 and 200 (if the patients were still being treated), were documented. For the first patch prescription, a single prescription of the smallest package (5 patches) was excluded if it was followed by a constant trend, such as stable, increasing, or decreasing dosages of at least 25 patches (eg, first prescription of 5 patches of TD fentanyl 25 pg/h, followed by 250 patches of TD fentanyl 50 lag/h). In these cases, the first dosage was considered to be a test prescription rather than the first effective dose. The overall time of patch application was documented in all patients. The treatment duration was calculated by multiplying the number of patches per patient by 3 (assuming an application time of 3 days per patch). If application was obviously double, the number of patches was counted as calculated for double application. For example, if a concomitant prescription of 5 patches of 100 lag/h + 5 patches of 50 pg/h was obviously meant for concomitant application,
the number of patches was counted as 5 patches (of 150 ~g/h).
analgesic coprescription did not follow, this could not be characterized as a true comedication.
Mean Cohort Analyses
Statistical Evaluations
Mean changes in dosages were analyzed for the cohort groups. The mean percentile increases were evab uated over the whole treatment period and the mean increases per day were calculated, taking into account the mean treatment duration.
Premedication and comedication were compared between TD fentanyl and TD buprenorphine patients using a X2 test. Increases in dosages were compared by paired Student t test. 3° P < 0.05 was considered statistically significant; values of P < 0.001 were additionally mentioned if applicable.
Intraindividual Analyses Intraindividual analyses were also performed and can be considered more accurate than the mean cohort analyses because they were generated on a per-patient basis, with mean values calculated only in the end. These analyses also allowed calculation of changes in dosages between the first prescription and day 100 as well as between day 100 and day 200. The percentile intraindividual increases per day (on a per-patient basis) were also determined, and the average values were calculated.
Analgesic Comedication For comparison of the dosages and changes in dosages during analgesic patch treatment, it was essential to assess the comedications prescribed as rescue medication to ensure that the prescribed patch treatment (ie, analgesic treatment) was similar between the TD fentanyl and TD buprenorphine patients and was not affected by concomitant analgesics of different strengths prescribed for breakthrough pain. Analgesic comedication was evaluated for the patients in the cancer and noncancer groups in the same manner as the evaluation of premedication (classification according to steps I, II, and III of the WHO ladder for pain treatment), and again following the methodology described in the previous study. 23 If no analgesics were prescribed concomitantly, this was classified as no comedication. Patients receiving analgesic comedications from different steps of the WHO ladder, such as steps I and III, were classified as belonging to the highest step (eg, step III in this example). A single prescription of an analgesic together with the first patch prescription was classified as no comedication if it was not followed by any further comedication. These single analgesic prescriptions were presumably given because of uncertain requirements for analgesia when starting the new patch treatment. If further
RESU LTS Patient Data In total, 448 patients with noncancer pain and 446 patients with cancer pain were included (552 women and 342 men). The mean age was 74 years (range, 25-101 years). Patient distribution and demographic data are shown in Table I. Previous analgesic medication (analgesic premeditation), as classified according to steps I, II, or III of the WHO ladder for pain treatment, 3 was comparable between the cohort groups. No significant differences in premedication between TD fentanyl and TD buprenorphine patients were found (Figure 1). Consequently, the patients were considered comparable with respect to pain intensity within the noncancer and cancer groups.
Changes in Dosages The mean prescribed dosages increased in all groups during treatment, but the increases were more pronounced in the TD fentanyl groups than in the TD buprenorphine groups.
Mean Cohort Analyses The mean percentile increases in dosages over the course of treatment were 32.5% for TD fentanyl and 14.8% for TD buprenorphine in noncancer patients, and 42.7% for TD fentanyl and 21.7% for TD buprenorphine in cancer patients. The differences between TD fentanyl and TD buprenorphine patients were statistically significant (P < 0.05). Treatment duration did not differ significantly between TD fentanyl and TD buprenorphine (Table II). From the mean increases over the whole treatment duration, the mean increases per day were calculated. The resulting increases were 0.11% for TD fentanyl and 0.07% for TD buprenorphine in the noncancer groups, and 0.18% for TD fentanyl and 0.10% for TD buprenorphine in the cancer groups (P < 0.05).
:Clfiltcal Theeap, eUtlCS>~-:......... :.;g:.. ~:,,,2. ~,,::: ,.'_L°7%~:T,:,:?V;:\,~:::,:-o~::0 %7°-2:::,7 ,~ -7 Z,TT::: ,;L::>:'r::~'~:::::;~:-:tx{{%::::,'--:,:-,.:,, ,)c
Table I. Distribution and demographic characteristics o f the patient groups. ~ Cancer Group
Noncancer Group TD Fentanyl
TD Buprenorphine
TD Fentanyl
TD Buprenorphine
Number
333
115
392
54
Age, y Mean Range
75 35-101
76 36-98
70 39-95
75 25-94
Sex, no. Female Male
242 91
196 196
29 25
85 30
TD = transdermal. *Both cohort groups received prescriptions for >30 TD fentanyl or TD buprenorphine patches; met International Statistical Classification of Diseases and Related Health Problems, l Oth Revision, 24 codes for cancer or noncancer and received similar analgesic premedication, which was classified according to steps I, II, and III of the World Health Organization analgesic ladder as a measure of" pain intensity.
[] [] [] []
70
7060so•~ 40 e_ 30t4.-
65.2
60 C
50
42.6
r-
40 t4-O
No premedication Step I Step II Step III
30
O
20
o~ 2010-
10 0
I
TD Fentanyl
I
TD Buprenorphine
Noncancer Patients
0-
r
TD Fentanyl
TD Buprenorphine
Cancer Patients
Figure 1. Distribution o f previous analgesic medication ("premedication") in patients with noncancer pain (n = 448) and cancer pain (n = 446) in the transdermal (TD) fentanyl and buprenorphine groups. Premedication was classified according to the World Health Organization ladder for pain treatment. 3 No significant differences were found between TD fentanyl and TD buprenorphine patients (X 2 test, I e > 0.05).
Intraindividual Analyses On an intraindividual basis, changes in dosages on fixed treatment days were estimated. Increases in the dosages between the first prescription and prescription on day 100 were calculated in noncancer patients to be 2 8 . 1 % for T D fentanyl and 9.4% for T D buprenorphine; corresponding values in cancer patients were 48.1% and 24.4%. Increases in the dosages between days 100
and 200 were lower, at 13.7% and 4 . 3 % in noncancer patients and 2 0 . 0 % and 0.8% in cancer patients for TD fentanyl and T D buprenorphine, respectively (Table III). However, the observed changes in dosages on fixed treatment days or differences between groups were not assessed to confirm statistical significance. O v e r the w h o l e course of t r e a t m e n t d u r a t i o n (dosages at first p r e s c r i p t i o n vs last prescription), the
Table II. Mean changes in dosages in the cohort groups.
- ..................
Noncancer Group
*~' [
I[
Cancer Group
I TD Fentanyl TD Buprenorphine TD Fentanyl TD Buprenorphine [ Mean Mean Mean Mean
treatment duration, d dosage on first prescription, lag/h dosage on last prescription, lag/h increase from first to last prescription, %
(n = 333)
(n = 115)
(n = 392)
(n = 54)
l
301 37.9 50.2
227* 39.9 45.8
234 47.3 67.5
225* 43.4 52.8
I I i
32.5
14.8f
42.7
21.7f
i I
I
TD = transdermal. *P > 0.05 versus TD fentanyl. fP < 0.05 versus TD fentanyl.
Table III. Changes in dosages on a per-patient basis (intraindividual increases). Noncancer Group
Cancer Group
T D Fentanyl
T D Buprenorphine
T D Fentanyl
TD Buprenorphine
Changes in dosages on first prescription vs day 100, % (n = patients available for calculation)
28.1 (n = 293)
9.4 (n = 94)
48.1
(n = 330)
24.4 (n = 45)
Changes in dosages on day 100 vs day 200, % (n = patients available for calculation)
13.7 (n = 173)
4.3 (n = 48)
20.0
0.8
(n = 155)
(n = 21)
48.1
17.2" (n = 115)
71.6 (n = 392)
25.9* (n = 54)
Mean intraindividual increases, first prescription vs last prescription, %
(n = total no. of patients in the cohort)
(n = 333)
TD = transdermal. *P < 0.001 versus TD fentanyl.
percentile intraindividual increases totaled 48.1% for TD fentanyl and 17.2% for TD buprenorphine in the noncancer groups, and 71.6% for TD fentanyl and 2 5 . 9 % for TD buprenorphine in the cancer groups. The higher increases in the dosages in TD fentanyl patients were statistically significant (P < 0.001; Table III). Taking into consideration the treatment durations, the mean percentile intraindividual increases per day were calculated to be 0.25% for TD fentanyl and 0.09% for TD buprenorphine in noncancer patients; corresponding values were 0.42% and 0.17% in cancer patients. These differences were statistically significant (P < 0.001; Figure 2). ~l~-~;~ll~ ~ ~ S ~ , ~'2,~,~
-~ ~ ~ z , ~ 2 ~
~
,
~--~,..........~ ~ 7 ~
Analgesic Comedication No significant differences in analgesic comedication were observed between the TD fentanyl and TD buprenorphine cohorts (Figure 3). DISCUSSION No comparison of dosage changes in fentanyl and buprenorphine analgesic patches over the course of treatments was found in literature searches. We used the database IMS Disease Analyzer-mediplus to evaluate changes in the prescribed dosages of TD fentanyl and TD buprenorphine in patients with cancer and noncancer pain in Germany. The database has proven useful in qualitative and quantitative analyses in health ,-~-o ~ . ~ - ~ w ~ ~ : - , ~ ~,~,, ~ , ~ 5 , ~ , ~
"~
........ ~ , ~ ~ %~:~
~
~
~-~-~~,,,,l~¢1'1~
5 ~ . ~ j ~ : ~"~~, ~.......... ~-~ . . . .
0.5--
0.5-
0.4-
0.4-
k)
0.25
¢" 0.30.2-
¢- 0.3 0.2
l
O
o~
U
0.10
0.09* i
TD Fentanyl
O 0.1 i
0
0.42
i
0.17"
I
TD Fentanyl
TD Buprenorphine
TD Buprenorphine
I
Cancer Patients
Noncancer Patients
Figure 2. Mean percentile intraindividual increases per day in transdermal ( T D ) buprenorphine and T D fentanyl dosages in noncancer and cancer patients. The differences between TD fentanyl and TD buprenorphine were statistically significant (Student t test, P < 0.001 ). Bars indicate standard error. *P < 0.001.
• N o comedication
[] [] []
45t
42.6 39.3
40
c 'n e_ o
23.4
o
40.7
4O
~3
q..
Step I Step II Step III
l
TD Fentanyl
I
TD Buprenorphine
35 302520-
31.9
151050-
I
TD Fentanyl
TD Buprenorphine
C a n c e r Patients Noncancer Patients Figure 3. Distribution ofcomedication in patients with noncancer pain (n = 448) and cancer pain (n = 446) in the transdermal (TD) fentanyl and buprenorphine groups. Comedication was classified according to the World Health Organization ladder for pain treatment. 3 No significant differences were found between TD fentanyl and TD buprenorphine patients (X2 test, P > 0.05).
care research 22 and specifically for retrospective dose
evaluations.23 Retrospective analyses in general have limitations because they are not double blind, randomized, or placebo controlled by definition. On the other hand, they allow researchers to examine medical care utilization as it occurs in routine clinical practice. 31 In addition, the data they produce are unbiased because neither patients nor physicians are aware of a study at the time prescriptions are dispensed. However, there is a risk of selection bias, particularly when defining subgroups. We attempted to avoid selection bias by performing the analyses as described previously, therefore following the methodology that was used in the previ-
ous study, where selection bias was shown to be irrelevant, because the results corresponded to those obtained in all patients. 23 A limitation of the database itself is that reasons for withdrawals were not documented. However, because the treatment duration did not differ significantly between the TD fentanyl and TD buprenorphine groups, there is no indication of earlier treatment termination in either of the groups, regardless of the reason. Because of another limitation of the database, the maximum detectable dosages were 175 pg/h for TD fentanyl and 122.5 pg/h for TD buprenorphine. This limitation led to a significant bias in the TD fentanyl group because no maximum dosage was stipulated
in the prescribing information. 27 Patients may apply much higher dosages of TD fentanyl than the maximum dosage detectable in our analysis, given that effective dosages for the treatment of cancer pain are up to 1000 t.~g/h28 and for noncancer pain are up to 325 gg/h. ~9 In contrast, underestimation of the buprenorphine dosages is less pronounced because the maximum dosage is 2 of the largest patches (140 l~g/h),26 which is only slightly above the detection threshold in these analyses. A concomitant application of >2 patches was not identifiable in the database. The application of >2 patches concomitantly is only recommended with TD fentanyl and has been well documented in the literature (eg, up to 10 patches). 28 Because the number of concomitantly applied buprenorphine patches is limited to 2, an underestimation of dosages was again less evident with these patches. In fact, we cannot rule out that physicians prescribed outside the label recommendations in Germany, resulting in >2 patches of TD buprenorphine. However, underestimation can still be assumed to be greater with TD fentanyl because multiple applications are not limited and are often reported. 4,10-12,28,29 The assessment of double application cannot be stated to be unbiased because the evaluations were not blinded to treatment. However, assessment was performed as described in the previous study, in which bias was found to be irrelevant due to matching results to groups without assumptions on double application. 23 Because of the underestimation of the TD fentanyl dosages, the differences in increases may presumably be even higher than those that we found. For the purposes of validation, we compared our results with the mean increases in dosages calculated from published studies on TD fentanyl showing the first and last prescribed dosages as well as treatment duration. Mean percentile daily increases with TD fentanyl were 0.47% to 3.38% for cancer pain ~°,~ and 0.24% to 0.53% for noncancer pain.~2, ~3 In our study, the mean daily increases for TD fentanyl in cancer patients totaled 0.18% and 0.42% for the mean cohort and for the intraindividual analyses, respectively; corresponding values for noncancer patients were 0.11% and 0.25%. In general, the intraindividual data may be considered more precise because they were generated from single patient data, from which mean values were subsequently calculated, whereas in the mean cohort analyses, changes in the dosage were calculated
.....
~, ~.~ .....
L~.Z ..... ~ " ~ ' ~
v.................
. . / ~ - ~ . ~ . / . . . . . . . . ~ ~,h " ~ ' ~ ~'~'÷~'~.'- ~
from mean values. The lower increases in cancer patients found in our retrospective study might be due to underestimation of the fentanyl dosages, as described earlier. A more likely reason, however, is that the populations included in previous clinical studies on TD fentanyl were hospital based and presumably were suffering from more severe diseases, whereas our study included subjects treated as outpatients by general practitioners or internists. No statistically significant difference was found in the treatment duration between the TD buprenorphine and TD fentanyl groups. However, the duration was slightly longer in the TD fentanyl noncancer patients. Therefore, the mean percentile increases up to treatment days 100 and 200 were also calculated to exclude an exponential increase in dosages. These calculations were performed on a per-patient (intraindividual) basis because some patients had dropped out by days 100 and 200. In both treatment groups, the percentile increases were higher up to day 100 than the increases between days 100 and 200, indicating that a dose-titration phase was still ongoing at the beginning of the analgesic treatment. This effect was not prevented by excluding a single prescription of 5 patches that were obviously meant for dose titration. However, an exponential dosage increase that might influence the differences in dosage increases in noncancer patients was ruled out by the additional calculations on fixed treatment days. We also analyzed comedication because the prescription of breakthrough pain medication in addition to a sustained-release basic medication is recommended by the W H O guidelines for cancer patients. 3 The prescription of additional analgesics is also often necessary in noncancer patients. In this study, analysis of comedication was restricted to analgesics belonging to the EPhMRA classifications N02A, N02B, and M 0 1 A Y Thus, nonanalgesic comedications, such as antidepressants, were not considered. This seemed appropriate because the indications for the prescription of those drugs (eg, depression, pain) could not be assessed easily. No significant difference was found in comedication between the TD fentanyl and TD buprenorphine cohorts. For this reason, it may be assumed that pain relief was similar with both patch treatments. Thus, there was no lower or more restricted dosing of TD buprenorphine that could explain the differences in dosage changes between the patches.
-5~;~
.......
~2.~
2 ~ % ' ~ . ~. . . . . . . . . .
~. . . . . . . . ~ - ~
~ 4 . . ~ . . . . . . . . . ~7~~ ' ~ "
>'~;r ......................
We compared changes in the TD fentanyl and TD buprenorphine dosages in cancer and noncancer patients. In these cohort patients, pain at baseline was considered to be similar because of comparable analgesic medication before beginning the patch treatment. Therefore, a comparable dosage development might have been expected. In the mean cohort and the intraindividual analyses, the mean percentile increases in dosages over the whole treatment and the increases per day were significantly higher in TD fentanyl patients than in TD buprenorphine patients. In all groups, increases in dosages during treatment were a result of dose titration, deterioration of the underlying disease, tolerance development, or a combination of these factors. The higher increases in dosages in the cancer groups (both treatments) than in the noncancer groups could be explained by progression of the cancer. However, deterioration of the underlying disease may not explain entirely the differences between the 2 analgesic patches within the cancer and noncancer groups because the analyses were conducted on cohort patients whose pain intensity at baseline was assumed to be similar, as described above. A titration phase was also expected to be similar in the cohort groups and did not explain the differences between TD fentanyl and TD buprenorphine. Finally, there was no indication of an inadequate starting dosage in the TD fentanyl groups that would explain the higher increase in dosages. Taking into account the equipotency ratios for TD fentany123,27 and TD buprenorphine, 23,26 there is no evidence of a lower starting dosage in the TD fentanyl groups. Our results suggest that tolerance development may be higher with TD fentanyl than with TD buprenorphine. Faster tolerance development may result in poorer pain control and more adverse effects (at least for those adverse effects not affected by tolerance development) as well as higher drug costs. However, these results must be confirmed in prospective, randomized, controlled studies. CONCLUSIONS These retrospective analyses were conducted on cohort patients in whom pain intensity was assumed to be similar because analgesic premedication was comparable. We found significantly higher increases in the mean daily doses in the TD fentanyl groups (noncancer and cancer pain) than in the TD buprenorphine groups.
ACKNOWLEDGMENTS This study was supported by Gruenenthal GmbH, Aachen, Germany. Dr. Poulsen Nautrup is an employee of Gruenenthal GmbH. Drs. Sittl and Nuijten were consultants for Gruenenthal GmbH. REFERENCES 1. Zech DF, Grond S, Lynch J, et al. Validation of World Health Organization Guidelines for cancer pain relief." A 10-year prospective study. Pain. 1995;63:65-76. 2. Schug SA, Merry AF, Acland RH. Treatment principles for the use ofopioids in pain of nonmalignant origin. Drugs. 1991;42:228-239. 3. WHO guidelines. CancerPainReliefand Pallia~iveCam. Geneva, Switzerland: World Health Organization; 1990. 4. Ackerman SJ,-Mordin M, Reblando J, et al. Patientreported utilization patterns of fentanyl transdermal system and oxycodone hydrochJoride controlled-release among patients with chronic nonmalignant pain.J Mana~ Care Pharm. 2003;9:223-231. 5. Mercadante S. Opioid rotation for cancer pain: Rationale and clinical aspects. Cancer. 1999;86:1856-1866. 6. Bailey CP, Connor M. Opioids: Cellular mechanisms of tolerance and physical dependence. Curt Opin Pharmacol. 2005;5:60-68. 7. Hsu MM, Wong CS. The roles of pain facilitatory systems in opioid tolerance. Ada AnaesthesiolSin. 2000;38:155-166. 8. Freye E, Latasch L. Development of opioid tolerancemolecular mechanisms and clinical consequences [in German]. Anas~hesiolln~ensivmedNot~llmedSchmerz~her.2003; 38:14-26. 9. Muijsers RB, WagstaffAJ. Transdermal fentanyl: An updated review of its pharmacological properties and therapeutic efficacy in chronic cancer pain control. Drugs. 2001 ;61:2289-2307. 10. Mystakidou K, Tsilika E, Parpa E, et al. Long-term cancer pain management in morphine pre-treated and opioid naive patients with transdermal fentanyl. In~JCancer. 2003; 107:486-492. 11. Korte W, de Stoutz N, Morant R. Day-to-day titration to initiate transdermal fentanyl in patients with cancer pain: Short- and long-term experiences in a prospective study of 39 patients.J Pain SymptomManage. 1996;11:139-146. 12. Mystakidou K, Parpa E, Tsilika E, et al. Long-term management of noncancer pain with transdermal therapeutic system-fentanyl. J Pain. 2003;4:298-306. 13. Milligan K, Lanteri-Minet M, Borchert K, et al. Evaluation of long-term efficacy and safety oftransdermal fentanyl in the treatment of chronic noncancer pain.J Pain. 2001 ;2: 197-204. 14. Heel RC, Brodgen RN, Speight TM, Avery GS. Buprenorphine: A review of its pharmacological properties and
15.
16.
17.
18.
19.
20.
21.
22.
23.
~
y
therapeutic efficacy. Drugs. 1979;17: 81-110. Finco G, Polati E, Bartoloni A, et al. Intravenous patient-controlled analgesia (PCA) in the treatment of postoperative pain: Rationale and clinical application [in Italian]. Chit Ital. 1995;74:20-25. B6hme K, Likar R. Efficacy and tolerability of a new opioid analgesic formulation, buprenorphine transdermal therapeutic system (TDS), in the treatment of patients with chronic pain. A randomised, double-blind, placebo-controlled study. Pain Clinic. 2003;15:193-202. Sittl R, Griessinger N, Likar R. Analgesic efficacy and tolerability oftransdermal buprenorphine in patients with inadequately controlled chronic pain related to cancer and other disorders: A multicenter, randomized, double-blind, placebo-controlled trial. Clin Ther. 2003;25:150-168. SorgeJ, Sittl R. Transdermal buprenorphine in the treatment of chronic pain: Results of a phase III, multicenter, randomized, double-blind, placebo-controlled study. Clin Ther. 2004;26:1808-1820. Allan L, Hays H, Jensen NH, et al. Randomised crossover trial oftransdermal fentanyl and sustained release oral morphine for treating chronic non-cancer pain. BMJ. 2001; 322:1154-1158. Payne R, Mathias SD, Pasta DJ, et al. Quality of life and cancer pain: Satisfaction and side effects with transdermal fentanyl versus oral morphine. J Clin Oncol. 1998;16:1588-1593. Robinson SE. Buprenorphine: An analgesic with an expanding role in the treatment of opioid addiction. CNS Drug Rev. 2002;8:377-390. Dietlein G, Schroder-Bernhardi D. Use of Mediplus patient database in healthcare research. IntJ Clin Pharmacol Ther. 2002;40:130-133. Sittl R, Likar R, Nautrup BP. Equipotent doses of transdermal fentanyl and transdermal buprenorphine in patients with cancer and
#
~
.
~
,
.
~
-
~
.......
~
~
a
~
~
~
.
~
noncancer pain: Results of a retrospective cohort study. Clin Ther. 2005; 27:225-237. 24. International Statistical Classification of Diseases and Related Health Problems.
28. Donner B, Zenz M, Strumpf M, Raber M. Long-term treatment of cancer pain with transdermal fentanyl. J Pain Symptom Manage. 1998; 15:168-175.
10th rev.Version for 2003. Available at: http://www.who.int/icd/vo[1 htm2003. Accessed November 19, 2004. 25. European Pharmaceutical Marketing Research Association, Anatomical Classification System. Available at: http://www.ephmra.org. Accessed November 19, 2004. 26. Transtec [summary of product characteristics]. Aachen, Germany: Gruenenthal GmbH; 2003. 27. Durogesic [packet insert]. Neuss, Germany: Janssen Pharmaceutical Products LP; 2001.
29. Radbrueh L, Sabatowski R, Petzke F, et al. Transdermal fentanyl for the management of cancer pain: A survey of1005 patients. PalliatMed. 2001 ;15:
309-32t. 30. Armitage P, Berry G. StatisticalMethods in Medical Research. Oxford, UK: Blackwell; 1994. 31. Motheral B, Brooks J, Clark MA, et al. A checklist for retrospective
database studies--Report of the ISPOR Task Force on Retrospective Databases. Value Health. 2003;6:9097.
Address correspondence to: Barbara Poulsen N a u t r u p , PhD, Strategic Marketing, H e a d of P h a r m a c o e c o n o m i c s , Gruenenthal G m b H , 5 2 0 9 9 Aachen, Germany. E-mail: b a r b a r a . p o u l s e n - n a u t r u p @ g r u n e n t h a l . d e :
~ f
. ~
:
W
,
.
~ ~
L .
.
~
-
~
,~ ........
~
9
~ ~ ~ ",~: ~z~ ~ :.~ =, ~" ~ . ~ ~ ~ ~
~.~9~
~ ~.~e~g-~e~#~.~,~,~:~-.~.
~-~ e
~~ ~
~:~¢~o~ "~ .
~L~ ~
.
! o J ~ : [ ~~. ~
-
~
e
•~ *
'~ ~ Ciinicai ~Therapeutics/Vo|urne27, Number ! 7, 20Q$ ~ :~ :
;~ i~:::
:~:i
' '~~
Changes in the Prescribed Daily Doses of Transdermal Fentanyl and Transdermal Buprenorphine During Treatment of Patients with Cancer and Noncancer Pain in Germany: Results of a Retrospective Cohort Study Reinhard Sittl, MD1; Marl< Nuijten, MD, PhD2; and Barbara Poulsen Nautrup, PhD 3
1Pain Clinic, Universityof Erlangen, Erlangen, Germany;2MEDTAPInternational, Amsterdam, the Netherlands; and 3Gruenentkal GmbH, Aachen, Germany ABSTRACT Background: During long-term therapy with strong opioids (step III opioids according to the World Health Organization [WHO] analgesic ladder), dose increases are often necessary because of deterioration of the primary disease or development of tolerance. Objective: The purpose of this study was to compare changes in dosages of transdermal (TD) fentanyl and TD buprenorphine in patients with cancer and noncancer pain. Methods: In a retrospective study, patients with cancer and noncancer pain being treated with TD fentanyl or TD buprenorphine for at least 3 months between January 2001 and December 2003 were identified from the IMS Disease Analyzer-mediplus database, which contains all patient-related data documented from 400 medical practices in Germany. The indications for treatment were define, t according to the International Classification of Diseases, l Oth Revision, and included neoplasm (cancer groups), and osteoarthritis, low back pain, and osteoporosis (noncancer groups). The cohort patients were considered to have comparable pain intensity because they had received similar analgesic premedication classified according to steps I to III of the WHO analgesic ladder (cohort groups). The mean prescribed daily doses on first and last prescription were documented, and the mean percentile increases were calculated over the whole treatment duration and per day. Additionally, the mean percentile intraindividual increases (on a per-patient basis) were estimated. Results: The cohort groups consisted of 448 patients with noncancer pain and 446 patients with cancer pain (552 women and 342 men; mean age, 74 years; range, 25-101 years). The mean percentile increases in dosages over the whole treatment duration and adjusted per day were significantly higher in patients
1022
., ,~,
taking TD fentanyl (P < 0.05). Differences were even greater for the mean percentile intraindividual increases per day, which totaled 0.42% and 0.17% for cancer patients taking TD fentanyl and TD buprenorphine, respectively; corresponding values were 0.25% and 0.09% in noncancer patients (P < 0.001). Conclusions: This retrospective analysis showed a significantly higher increase in the mean daily doses of TD fentanyl as compared with TD buprenorphine. The results must be verified in prospective, randomized clinical studies. (Clin Ther. 2005;27:1022-1031) Copyright © 2005 Excerpta Medica, Inc.) Key words: analgesic, buprenorphine, tolerance, fentanyl, transdermal.
INTRODUCTION Opioids are the mainstay of treatment for chronic moderate to severe pain. 1,2 The goal of pain management with opioids is to provide optimal analgesia with a minimum of adverse effects. The World Health Organization (WHO) guidelines on cancer pain management stipulate adapting the dosage of analgesics to the degree of pain. 3 Accordingly, there are no standard doses for opioid drugs. The "right" dose is the dose that relieves the patient's pain. 3 This work was presented in part as an abstract at the International Society for Pharmacoeconomics and Outcomes Research 10th Annual International Meeting, May 15-18, 2005, Washington, DC.
Accepted fiorpublication June 2, 2005. Express Track online publication June 28, 2005. doi:l 0.1016/j.clinthera.2005.06.024 0149-2918/05/$19.00 Printed in the USA. Reproduction in whole or part is not permitted. Copyright © 2005 Excerpta Medica, Inc.
~
~
. .
.
.
.
.
Volume-27 Number 7 .
During long-term therapy with strong opioids, dose increases are often necessary because of deterioration of the primary disease, development of tolerance, or both. 4 Analgesic tolerance is defined pharmacologically as a reduced analgesic effect of opioids after repeated administration or the need for higher doses to maintain the same effect. 5 Tolerance develops over hours, days, or weeks of opioid application. 6 Opioid tolerance mainly develops on the level of the target cells; 2 well-described mechanisms include opioid receptor uncoupling from G-proteins and receptor downregulation. 7 Clinically, using high dosages of opioids to overcome tolerance may result in a greater risk of adverse effects because of an increase in toxic opioid metabolites, s Tolerance to the opioid's effects (therapeuuc and toxic) develops at different rates. Among the adverse effects, the development of tolerance is mimmal for constipation and miosis, whereas tolerance occurs rapidly to nausea, vomiting, sedation, euphoria, and respiratory depression, s For chronic pain treatment, the transdermal (TD) opioid formulations of fentanyl* and buprenorphine~ have been used in many countries. Fentanyl is a full agomst at p-opioid receptors and was introduced into medical practice in the 1960s. It is a lipid-soluble synthetic opioid that can be delivered in a TD controlledrelease formulation for up to 72 hours. 9 TD fentanyl has proved effective in the treatment of cancer and noncancer pain. ~°-~3 TD buprenorphine is a centrally acting analgesic that was introduced to European markets >20 years ago and is some 25 to 50 times more potent than morphine on a weight-by-weight basis34,~s The compound has a high affinity for opioid receptors, acting as a partial agonist at la-opioid receptors and as an antagonist at ~:-opioid receptors. Beginning in 2001, the TD buprenorphine formulation has been launched in many European countries. The efficacy and safety of TD buprenorphine have been demonstrated in randomized, placebo-controlled studies in cancer and noncancer patients. 16-~8 TD formulations are new treatment options for longterm pain management in patients who require stable dosing. The advantages of TD therapeutic systems include maintenance of constant plasma levels, which have a favorable impact on the incidence of adverse ef~Trademark: Durogesic® (Janssen Pharmaceutical Products kP, Neuss, Germany). 1"Trademark: Transtec® (Gr0nenthal GmbH, Aachen, Germany).
fects. Some evidence indicates that TD opioids may improve patients' compliance and quality of life during long-term therapy39,2° For TD fentanyl, cancer pain studies have documented significant increases in the dosages during treatment. Mean initial versus mean final dosages were 38.7 versus 239.6 pg/h (mean treatment duration, 1095 days), ~° respectively, and 44.0 versus 178.0 lag& (mean treatment duration, 90 days), l~ In studies of noncancer pain, dosage increases were also reported from 25.8 to 67.3 lag& (mean treatment duration, 304 days) 12 and from 48.0 to 90.0 tlg/h (treatment duration, 365 days). 13 However, these studies did not differentiate the reasons for increased dosages, whether dose titration, deterioration of the underlying disease, or tolerance development. With buprenorphine treatment, tolerance development has been reported to be low, but these data were not based on the TD bnprenorphine formulation, and no figures on tolerance development were reported34,2~ Literature searches in MEDLINE and EMBASE for the years 1995 through 2005 did not find any study on TD buprenorphine that documented the initial and final dosages (ie, changes in dosages during treatment). Accordingly, we did not find a direct comparison in the literature of dosage changes between TD fentanyl and TD buprenorphine in comparable patients according to pain and underlying disease. The aim o f this retrospective study was to evaluate the prescribed dosages of TD fentanyl and TD buprenorphine that were documented over a period of 3 years in Germany in patients with cancer and noncancer pain (due to osteoarthritis, low back pain, and osteoporosis) and to compare the dosage changes over the course of treatment. PATIENTS A N D METHODS Patient Data
Patient data analyzed in this study were derived retrospectively from the IMS Disease Analyzer-mediplus database, which has been used to investigate disease courses and therapy patterns. 22,23The database contains anonymous information on patients treated in 400 practices of general practitioners and internists in Germany. Patients'being treated with TD fentanyl or TD buprenorphine were analyzed for a documentation period of 3 years, between January 2001 and December 2003. Germany does not require institutional review board waiver of consent for retrospective studies of anonymous data.
Pain treatment was evaluated separately for noncancer and cancer patients. A similar methodology was used as described in a previous study, which analyzed the same database over a similar period of 3 years; however, the previous study covered a different time frame and had a different objective. 23 The indications for treatment were defined according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-1 O) diagnostic codes.24 The following indications were included: for the cancer groups, patients with the ICD-IO codes COO through C97 (neoplasm), and for the noncancer groups, patients with the ICD-10 codes M15 through M19 (osteoarthritis), M40 through M54 (low back pain), and M80 through M85 (osteoporosis). For a direct comparison of the dosages and dosage changes, it was essential to perform the analyses on comparable patients according to pain intensity. Because the analgesic dosage depends on the intensity of pain and is administered and titrated freely within the recommended dose range of the opioid, a comparison was rational only in patients with similar pain intensity. Because-pain intensity was not documented in the database used, patients were assumed to have comparable pain intensity if they had received similar analgesic medication before the patch treatment, as described in detail in a previous study.23 According to the methodology in this study, we defined a subgroup of patients on long-term TD fentanyl or TD buprenorphine treatment as _>3 months, which corresponds to >-30 patch prescriptions, assuming a wearing time of 3 days per patch as recommended. For evaluation of previously received analgesic medication (premedication), all drugs from the European Pharmaceutical Market Research Association (EPhMRA) classifications 2s of N02A (opioids), N02B (other analgesics), and M01A (anti-inflammatory and antirheumatic products) were considered and classified according to step I, II, or III of the WHO ladder for pain treatment.3 The WHO ladder specifies that the first step is a nonopioid, the second is a weak opioid for mild to moderate pain, and the third is a strong opioid for moderate to severe pain. If no analgesics were prescribed previously, this was classified as no premedication. Patients receiving premedication analgesics from different steps of the WHO ladder, such as steps I and III, were classified as belonging to the highest step (eg, step III in the given example).
1024::
Assessmentof Dosages The daily doses were determined directly from the prescribed patch strength. However, when calculating the prescribed daily doses, we considered that because of the label recommendations, patients could apply > 1 patch at a time,26,27 a fact that has often been reported for TD fentanyl.4,1°-12,28,29 However, the database used did not explicitly mention when 2 patches had to be applied concomitantly. Regarding concomitant prescriptions of different patch strengths, double patch applications were taken into consideration only if they were obvious. For example, a concomitant prescription of 100 + 50 lag/h of TD fentanyl was only considered to be a double application if the resulting dosage seemed rational with regard to the previously and subsequently prescribed doses. If a concomitant prescription on 1 day was obviously intended for dose titration, the double prescription was not counted as a concomitant application. The possibility of identifying patients who applied >1 patch was limited for daily doses >175 lag& for TD fentanyl (eg, 2 × 100 pg/h) or >122.5 lag& for TD buprenorphine. It was not possible to identify patients who received prescriptions of ] patch strength but applied _>2 patches (eg, 2 × 50 pg/h).
Changes in Dosages In all patients in the cancer and noncancer cohorts, the dosages prescribed at first and last prescription (regardless of the time until last prescription), as well as the dosages on days 100 and 200 (if the patients were still being treated), were documented. For the first patch prescription, a single prescription of the smallest package (5 patches) was excluded if it was followed by a constant trend, such as stable, increasing, or decreasing dosages of at least 25 patches (eg, first prescription of 5 patches of TD fentanyl 25 pg/h, followed by 250 patches of TD fentanyl 50 lag/h). In these cases, the first dosage was considered to be a test prescription rather than the first effective dose. The overall time of patch application was documented in all patients. The treatment duration was calculated by multiplying the number of patches per patient by 3 (assuming an application time of 3 days per patch). If application was obviously double, the number of patches was counted as calculated for double application. For example, if a concomitant prescription of 5 patches of 100 lag/h + 5 patches of 50 pg/h was obviously meant for concomitant application,
the number of patches was counted as 5 patches (of 150 ~g/h).
analgesic coprescription did not follow, this could not be characterized as a true comedication.
Mean Cohort Analyses
Statistical Evaluations
Mean changes in dosages were analyzed for the cohort groups. The mean percentile increases were evab uated over the whole treatment period and the mean increases per day were calculated, taking into account the mean treatment duration.
Premedication and comedication were compared between TD fentanyl and TD buprenorphine patients using a X2 test. Increases in dosages were compared by paired Student t test. 3° P < 0.05 was considered statistically significant; values of P < 0.001 were additionally mentioned if applicable.
Intraindividual Analyses Intraindividual analyses were also performed and can be considered more accurate than the mean cohort analyses because they were generated on a per-patient basis, with mean values calculated only in the end. These analyses also allowed calculation of changes in dosages between the first prescription and day 100 as well as between day 100 and day 200. The percentile intraindividual increases per day (on a per-patient basis) were also determined, and the average values were calculated.
Analgesic Comedication For comparison of the dosages and changes in dosages during analgesic patch treatment, it was essential to assess the comedications prescribed as rescue medication to ensure that the prescribed patch treatment (ie, analgesic treatment) was similar between the TD fentanyl and TD buprenorphine patients and was not affected by concomitant analgesics of different strengths prescribed for breakthrough pain. Analgesic comedication was evaluated for the patients in the cancer and noncancer groups in the same manner as the evaluation of premedication (classification according to steps I, II, and III of the WHO ladder for pain treatment), and again following the methodology described in the previous study. 23 If no analgesics were prescribed concomitantly, this was classified as no comedication. Patients receiving analgesic comedications from different steps of the WHO ladder, such as steps I and III, were classified as belonging to the highest step (eg, step III in this example). A single prescription of an analgesic together with the first patch prescription was classified as no comedication if it was not followed by any further comedication. These single analgesic prescriptions were presumably given because of uncertain requirements for analgesia when starting the new patch treatment. If further
RESU LTS Patient Data In total, 448 patients with noncancer pain and 446 patients with cancer pain were included (552 women and 342 men). The mean age was 74 years (range, 25-101 years). Patient distribution and demographic data are shown in Table I. Previous analgesic medication (analgesic premeditation), as classified according to steps I, II, or III of the WHO ladder for pain treatment, 3 was comparable between the cohort groups. No significant differences in premedication between TD fentanyl and TD buprenorphine patients were found (Figure 1). Consequently, the patients were considered comparable with respect to pain intensity within the noncancer and cancer groups.
Changes in Dosages The mean prescribed dosages increased in all groups during treatment, but the increases were more pronounced in the TD fentanyl groups than in the TD buprenorphine groups.
Mean Cohort Analyses The mean percentile increases in dosages over the course of treatment were 32.5% for TD fentanyl and 14.8% for TD buprenorphine in noncancer patients, and 42.7% for TD fentanyl and 21.7% for TD buprenorphine in cancer patients. The differences between TD fentanyl and TD buprenorphine patients were statistically significant (P < 0.05). Treatment duration did not differ significantly between TD fentanyl and TD buprenorphine (Table II). From the mean increases over the whole treatment duration, the mean increases per day were calculated. The resulting increases were 0.11% for TD fentanyl and 0.07% for TD buprenorphine in the noncancer groups, and 0.18% for TD fentanyl and 0.10% for TD buprenorphine in the cancer groups (P < 0.05).
:Clfiltcal Theeap, eUtlCS>~-:......... :.;g:.. ~:,,,2. ~,,::: ,.'_L°7%~:T,:,:?V;:\,~:::,:-o~::0 %7°-2:::,7 ,~ -7 Z,TT::: ,;L::>:'r::~'~:::::;~:-:tx{{%::::,'--:,:-,.:,, ,)c
Table I. Distribution and demographic characteristics o f the patient groups. ~ Cancer Group
Noncancer Group TD Fentanyl
TD Buprenorphine
TD Fentanyl
TD Buprenorphine
Number
333
115
392
54
Age, y Mean Range
75 35-101
76 36-98
70 39-95
75 25-94
Sex, no. Female Male
242 91
196 196
29 25
85 30
TD = transdermal. *Both cohort groups received prescriptions for >30 TD fentanyl or TD buprenorphine patches; met International Statistical Classification of Diseases and Related Health Problems, l Oth Revision, 24 codes for cancer or noncancer and received similar analgesic premedication, which was classified according to steps I, II, and III of the World Health Organization analgesic ladder as a measure of" pain intensity.
[] [] [] []
70
7060so•~ 40 e_ 30t4.-
65.2
60 C
50
42.6
r-
40 t4-O
No premedication Step I Step II Step III
30
O
20
o~ 2010-
10 0
I
TD Fentanyl
I
TD Buprenorphine
Noncancer Patients
0-
r
TD Fentanyl
TD Buprenorphine
Cancer Patients
Figure 1. Distribution o f previous analgesic medication ("premedication") in patients with noncancer pain (n = 448) and cancer pain (n = 446) in the transdermal (TD) fentanyl and buprenorphine groups. Premedication was classified according to the World Health Organization ladder for pain treatment. 3 No significant differences were found between TD fentanyl and TD buprenorphine patients (X 2 test, I e > 0.05).
Intraindividual Analyses On an intraindividual basis, changes in dosages on fixed treatment days were estimated. Increases in the dosages between the first prescription and prescription on day 100 were calculated in noncancer patients to be 2 8 . 1 % for T D fentanyl and 9.4% for T D buprenorphine; corresponding values in cancer patients were 48.1% and 24.4%. Increases in the dosages between days 100
and 200 were lower, at 13.7% and 4 . 3 % in noncancer patients and 2 0 . 0 % and 0.8% in cancer patients for TD fentanyl and T D buprenorphine, respectively (Table III). However, the observed changes in dosages on fixed treatment days or differences between groups were not assessed to confirm statistical significance. O v e r the w h o l e course of t r e a t m e n t d u r a t i o n (dosages at first p r e s c r i p t i o n vs last prescription), the
Table II. Mean changes in dosages in the cohort groups.
- ..................
Noncancer Group
*~' [
I[
Cancer Group
I TD Fentanyl TD Buprenorphine TD Fentanyl TD Buprenorphine [ Mean Mean Mean Mean
treatment duration, d dosage on first prescription, lag/h dosage on last prescription, lag/h increase from first to last prescription, %
(n = 333)
(n = 115)
(n = 392)
(n = 54)
l
301 37.9 50.2
227* 39.9 45.8
234 47.3 67.5
225* 43.4 52.8
I I i
32.5
14.8f
42.7
21.7f
i I
I
TD = transdermal. *P > 0.05 versus TD fentanyl. fP < 0.05 versus TD fentanyl.
Table III. Changes in dosages on a per-patient basis (intraindividual increases). Noncancer Group
Cancer Group
T D Fentanyl
T D Buprenorphine
T D Fentanyl
TD Buprenorphine
Changes in dosages on first prescription vs day 100, % (n = patients available for calculation)
28.1 (n = 293)
9.4 (n = 94)
48.1
(n = 330)
24.4 (n = 45)
Changes in dosages on day 100 vs day 200, % (n = patients available for calculation)
13.7 (n = 173)
4.3 (n = 48)
20.0
0.8
(n = 155)
(n = 21)
48.1
17.2" (n = 115)
71.6 (n = 392)
25.9* (n = 54)
Mean intraindividual increases, first prescription vs last prescription, %
(n = total no. of patients in the cohort)
(n = 333)
TD = transdermal. *P < 0.001 versus TD fentanyl.
percentile intraindividual increases totaled 48.1% for TD fentanyl and 17.2% for TD buprenorphine in the noncancer groups, and 71.6% for TD fentanyl and 2 5 . 9 % for TD buprenorphine in the cancer groups. The higher increases in the dosages in TD fentanyl patients were statistically significant (P < 0.001; Table III). Taking into consideration the treatment durations, the mean percentile intraindividual increases per day were calculated to be 0.25% for TD fentanyl and 0.09% for TD buprenorphine in noncancer patients; corresponding values were 0.42% and 0.17% in cancer patients. These differences were statistically significant (P < 0.001; Figure 2). ~l~-~;~ll~ ~ ~ S ~ , ~'2,~,~
-~ ~ ~ z , ~ 2 ~
~
,
~--~,..........~ ~ 7 ~
Analgesic Comedication No significant differences in analgesic comedication were observed between the TD fentanyl and TD buprenorphine cohorts (Figure 3). DISCUSSION No comparison of dosage changes in fentanyl and buprenorphine analgesic patches over the course of treatments was found in literature searches. We used the database IMS Disease Analyzer-mediplus to evaluate changes in the prescribed dosages of TD fentanyl and TD buprenorphine in patients with cancer and noncancer pain in Germany. The database has proven useful in qualitative and quantitative analyses in health ,-~-o ~ . ~ - ~ w ~ ~ : - , ~ ~,~,, ~ , ~ 5 , ~ , ~
"~
........ ~ , ~ ~ %~:~
~
~
~-~-~~,,,,l~¢1'1~
5 ~ . ~ j ~ : ~"~~, ~.......... ~-~ . . . .
0.5--
0.5-
0.4-
0.4-
k)
0.25
¢" 0.30.2-
¢- 0.3 0.2
l
O
o~
U
0.10
0.09* i
TD Fentanyl
O 0.1 i
0
0.42
i
0.17"
I
TD Fentanyl
TD Buprenorphine
TD Buprenorphine
I
Cancer Patients
Noncancer Patients
Figure 2. Mean percentile intraindividual increases per day in transdermal ( T D ) buprenorphine and T D fentanyl dosages in noncancer and cancer patients. The differences between TD fentanyl and TD buprenorphine were statistically significant (Student t test, P < 0.001 ). Bars indicate standard error. *P < 0.001.
• N o comedication
[] [] []
45t
42.6 39.3
40
c 'n e_ o
23.4
o
40.7
4O
~3
q..
Step I Step II Step III
l
TD Fentanyl
I
TD Buprenorphine
35 302520-
31.9
151050-
I
TD Fentanyl
TD Buprenorphine
C a n c e r Patients Noncancer Patients Figure 3. Distribution ofcomedication in patients with noncancer pain (n = 448) and cancer pain (n = 446) in the transdermal (TD) fentanyl and buprenorphine groups. Comedication was classified according to the World Health Organization ladder for pain treatment. 3 No significant differences were found between TD fentanyl and TD buprenorphine patients (X2 test, P > 0.05).
care research 22 and specifically for retrospective dose
evaluations.23 Retrospective analyses in general have limitations because they are not double blind, randomized, or placebo controlled by definition. On the other hand, they allow researchers to examine medical care utilization as it occurs in routine clinical practice. 31 In addition, the data they produce are unbiased because neither patients nor physicians are aware of a study at the time prescriptions are dispensed. However, there is a risk of selection bias, particularly when defining subgroups. We attempted to avoid selection bias by performing the analyses as described previously, therefore following the methodology that was used in the previ-
ous study, where selection bias was shown to be irrelevant, because the results corresponded to those obtained in all patients. 23 A limitation of the database itself is that reasons for withdrawals were not documented. However, because the treatment duration did not differ significantly between the TD fentanyl and TD buprenorphine groups, there is no indication of earlier treatment termination in either of the groups, regardless of the reason. Because of another limitation of the database, the maximum detectable dosages were 175 pg/h for TD fentanyl and 122.5 pg/h for TD buprenorphine. This limitation led to a significant bias in the TD fentanyl group because no maximum dosage was stipulated
in the prescribing information. 27 Patients may apply much higher dosages of TD fentanyl than the maximum dosage detectable in our analysis, given that effective dosages for the treatment of cancer pain are up to 1000 t.~g/h28 and for noncancer pain are up to 325 gg/h. ~9 In contrast, underestimation of the buprenorphine dosages is less pronounced because the maximum dosage is 2 of the largest patches (140 l~g/h),26 which is only slightly above the detection threshold in these analyses. A concomitant application of >2 patches was not identifiable in the database. The application of >2 patches concomitantly is only recommended with TD fentanyl and has been well documented in the literature (eg, up to 10 patches). 28 Because the number of concomitantly applied buprenorphine patches is limited to 2, an underestimation of dosages was again less evident with these patches. In fact, we cannot rule out that physicians prescribed outside the label recommendations in Germany, resulting in >2 patches of TD buprenorphine. However, underestimation can still be assumed to be greater with TD fentanyl because multiple applications are not limited and are often reported. 4,10-12,28,29 The assessment of double application cannot be stated to be unbiased because the evaluations were not blinded to treatment. However, assessment was performed as described in the previous study, in which bias was found to be irrelevant due to matching results to groups without assumptions on double application. 23 Because of the underestimation of the TD fentanyl dosages, the differences in increases may presumably be even higher than those that we found. For the purposes of validation, we compared our results with the mean increases in dosages calculated from published studies on TD fentanyl showing the first and last prescribed dosages as well as treatment duration. Mean percentile daily increases with TD fentanyl were 0.47% to 3.38% for cancer pain ~°,~ and 0.24% to 0.53% for noncancer pain.~2, ~3 In our study, the mean daily increases for TD fentanyl in cancer patients totaled 0.18% and 0.42% for the mean cohort and for the intraindividual analyses, respectively; corresponding values for noncancer patients were 0.11% and 0.25%. In general, the intraindividual data may be considered more precise because they were generated from single patient data, from which mean values were subsequently calculated, whereas in the mean cohort analyses, changes in the dosage were calculated
.....
~, ~.~ .....
L~.Z ..... ~ " ~ ' ~
v.................
. . / ~ - ~ . ~ . / . . . . . . . . ~ ~,h " ~ ' ~ ~'~'÷~'~.'- ~
from mean values. The lower increases in cancer patients found in our retrospective study might be due to underestimation of the fentanyl dosages, as described earlier. A more likely reason, however, is that the populations included in previous clinical studies on TD fentanyl were hospital based and presumably were suffering from more severe diseases, whereas our study included subjects treated as outpatients by general practitioners or internists. No statistically significant difference was found in the treatment duration between the TD buprenorphine and TD fentanyl groups. However, the duration was slightly longer in the TD fentanyl noncancer patients. Therefore, the mean percentile increases up to treatment days 100 and 200 were also calculated to exclude an exponential increase in dosages. These calculations were performed on a per-patient (intraindividual) basis because some patients had dropped out by days 100 and 200. In both treatment groups, the percentile increases were higher up to day 100 than the increases between days 100 and 200, indicating that a dose-titration phase was still ongoing at the beginning of the analgesic treatment. This effect was not prevented by excluding a single prescription of 5 patches that were obviously meant for dose titration. However, an exponential dosage increase that might influence the differences in dosage increases in noncancer patients was ruled out by the additional calculations on fixed treatment days. We also analyzed comedication because the prescription of breakthrough pain medication in addition to a sustained-release basic medication is recommended by the W H O guidelines for cancer patients. 3 The prescription of additional analgesics is also often necessary in noncancer patients. In this study, analysis of comedication was restricted to analgesics belonging to the EPhMRA classifications N02A, N02B, and M 0 1 A Y Thus, nonanalgesic comedications, such as antidepressants, were not considered. This seemed appropriate because the indications for the prescription of those drugs (eg, depression, pain) could not be assessed easily. No significant difference was found in comedication between the TD fentanyl and TD buprenorphine cohorts. For this reason, it may be assumed that pain relief was similar with both patch treatments. Thus, there was no lower or more restricted dosing of TD buprenorphine that could explain the differences in dosage changes between the patches.
-5~;~
.......
~2.~
2 ~ % ' ~ . ~. . . . . . . . . .
~. . . . . . . . ~ - ~
~ 4 . . ~ . . . . . . . . . ~7~~ ' ~ "
>'~;r ......................
We compared changes in the TD fentanyl and TD buprenorphine dosages in cancer and noncancer patients. In these cohort patients, pain at baseline was considered to be similar because of comparable analgesic medication before beginning the patch treatment. Therefore, a comparable dosage development might have been expected. In the mean cohort and the intraindividual analyses, the mean percentile increases in dosages over the whole treatment and the increases per day were significantly higher in TD fentanyl patients than in TD buprenorphine patients. In all groups, increases in dosages during treatment were a result of dose titration, deterioration of the underlying disease, tolerance development, or a combination of these factors. The higher increases in dosages in the cancer groups (both treatments) than in the noncancer groups could be explained by progression of the cancer. However, deterioration of the underlying disease may not explain entirely the differences between the 2 analgesic patches within the cancer and noncancer groups because the analyses were conducted on cohort patients whose pain intensity at baseline was assumed to be similar, as described above. A titration phase was also expected to be similar in the cohort groups and did not explain the differences between TD fentanyl and TD buprenorphine. Finally, there was no indication of an inadequate starting dosage in the TD fentanyl groups that would explain the higher increase in dosages. Taking into account the equipotency ratios for TD fentany123,27 and TD buprenorphine, 23,26 there is no evidence of a lower starting dosage in the TD fentanyl groups. Our results suggest that tolerance development may be higher with TD fentanyl than with TD buprenorphine. Faster tolerance development may result in poorer pain control and more adverse effects (at least for those adverse effects not affected by tolerance development) as well as higher drug costs. However, these results must be confirmed in prospective, randomized, controlled studies. CONCLUSIONS These retrospective analyses were conducted on cohort patients in whom pain intensity was assumed to be similar because analgesic premedication was comparable. We found significantly higher increases in the mean daily doses in the TD fentanyl groups (noncancer and cancer pain) than in the TD buprenorphine groups.
ACKNOWLEDGMENTS This study was supported by Gruenenthal GmbH, Aachen, Germany. Dr. Poulsen Nautrup is an employee of Gruenenthal GmbH. Drs. Sittl and Nuijten were consultants for Gruenenthal GmbH. REFERENCES 1. Zech DF, Grond S, Lynch J, et al. Validation of World Health Organization Guidelines for cancer pain relief." A 10-year prospective study. Pain. 1995;63:65-76. 2. Schug SA, Merry AF, Acland RH. Treatment principles for the use ofopioids in pain of nonmalignant origin. Drugs. 1991;42:228-239. 3. WHO guidelines. CancerPainReliefand Pallia~iveCam. Geneva, Switzerland: World Health Organization; 1990. 4. Ackerman SJ,-Mordin M, Reblando J, et al. Patientreported utilization patterns of fentanyl transdermal system and oxycodone hydrochJoride controlled-release among patients with chronic nonmalignant pain.J Mana~ Care Pharm. 2003;9:223-231. 5. Mercadante S. Opioid rotation for cancer pain: Rationale and clinical aspects. Cancer. 1999;86:1856-1866. 6. Bailey CP, Connor M. Opioids: Cellular mechanisms of tolerance and physical dependence. Curt Opin Pharmacol. 2005;5:60-68. 7. Hsu MM, Wong CS. The roles of pain facilitatory systems in opioid tolerance. Ada AnaesthesiolSin. 2000;38:155-166. 8. Freye E, Latasch L. Development of opioid tolerancemolecular mechanisms and clinical consequences [in German]. Anas~hesiolln~ensivmedNot~llmedSchmerz~her.2003; 38:14-26. 9. Muijsers RB, WagstaffAJ. Transdermal fentanyl: An updated review of its pharmacological properties and therapeutic efficacy in chronic cancer pain control. Drugs. 2001 ;61:2289-2307. 10. Mystakidou K, Tsilika E, Parpa E, et al. Long-term cancer pain management in morphine pre-treated and opioid naive patients with transdermal fentanyl. In~JCancer. 2003; 107:486-492. 11. Korte W, de Stoutz N, Morant R. Day-to-day titration to initiate transdermal fentanyl in patients with cancer pain: Short- and long-term experiences in a prospective study of 39 patients.J Pain SymptomManage. 1996;11:139-146. 12. Mystakidou K, Parpa E, Tsilika E, et al. Long-term management of noncancer pain with transdermal therapeutic system-fentanyl. J Pain. 2003;4:298-306. 13. Milligan K, Lanteri-Minet M, Borchert K, et al. Evaluation of long-term efficacy and safety oftransdermal fentanyl in the treatment of chronic noncancer pain.J Pain. 2001 ;2: 197-204. 14. Heel RC, Brodgen RN, Speight TM, Avery GS. Buprenorphine: A review of its pharmacological properties and
15.
16.
17.
18.
19.
20.
21.
22.
23.
~
y
therapeutic efficacy. Drugs. 1979;17: 81-110. Finco G, Polati E, Bartoloni A, et al. Intravenous patient-controlled analgesia (PCA) in the treatment of postoperative pain: Rationale and clinical application [in Italian]. Chit Ital. 1995;74:20-25. B6hme K, Likar R. Efficacy and tolerability of a new opioid analgesic formulation, buprenorphine transdermal therapeutic system (TDS), in the treatment of patients with chronic pain. A randomised, double-blind, placebo-controlled study. Pain Clinic. 2003;15:193-202. Sittl R, Griessinger N, Likar R. Analgesic efficacy and tolerability oftransdermal buprenorphine in patients with inadequately controlled chronic pain related to cancer and other disorders: A multicenter, randomized, double-blind, placebo-controlled trial. Clin Ther. 2003;25:150-168. SorgeJ, Sittl R. Transdermal buprenorphine in the treatment of chronic pain: Results of a phase III, multicenter, randomized, double-blind, placebo-controlled study. Clin Ther. 2004;26:1808-1820. Allan L, Hays H, Jensen NH, et al. Randomised crossover trial oftransdermal fentanyl and sustained release oral morphine for treating chronic non-cancer pain. BMJ. 2001; 322:1154-1158. Payne R, Mathias SD, Pasta DJ, et al. Quality of life and cancer pain: Satisfaction and side effects with transdermal fentanyl versus oral morphine. J Clin Oncol. 1998;16:1588-1593. Robinson SE. Buprenorphine: An analgesic with an expanding role in the treatment of opioid addiction. CNS Drug Rev. 2002;8:377-390. Dietlein G, Schroder-Bernhardi D. Use of Mediplus patient database in healthcare research. IntJ Clin Pharmacol Ther. 2002;40:130-133. Sittl R, Likar R, Nautrup BP. Equipotent doses of transdermal fentanyl and transdermal buprenorphine in patients with cancer and
#
~
.
~
,
.
~
-
~
.......
~
~
a
~
~
~
.
~
noncancer pain: Results of a retrospective cohort study. Clin Ther. 2005; 27:225-237. 24. International Statistical Classification of Diseases and Related Health Problems.
28. Donner B, Zenz M, Strumpf M, Raber M. Long-term treatment of cancer pain with transdermal fentanyl. J Pain Symptom Manage. 1998; 15:168-175.
10th rev.Version for 2003. Available at: http://www.who.int/icd/vo[1 htm2003. Accessed November 19, 2004. 25. European Pharmaceutical Marketing Research Association, Anatomical Classification System. Available at: http://www.ephmra.org. Accessed November 19, 2004. 26. Transtec [summary of product characteristics]. Aachen, Germany: Gruenenthal GmbH; 2003. 27. Durogesic [packet insert]. Neuss, Germany: Janssen Pharmaceutical Products LP; 2001.
29. Radbrueh L, Sabatowski R, Petzke F, et al. Transdermal fentanyl for the management of cancer pain: A survey of1005 patients. PalliatMed. 2001 ;15:
309-32t. 30. Armitage P, Berry G. StatisticalMethods in Medical Research. Oxford, UK: Blackwell; 1994. 31. Motheral B, Brooks J, Clark MA, et al. A checklist for retrospective
database studies--Report of the ISPOR Task Force on Retrospective Databases. Value Health. 2003;6:9097.
Address correspondence to: Barbara Poulsen N a u t r u p , PhD, Strategic Marketing, H e a d of P h a r m a c o e c o n o m i c s , Gruenenthal G m b H , 5 2 0 9 9 Aachen, Germany. E-mail: b a r b a r a . p o u l s e n - n a u t r u p @ g r u n e n t h a l . d e :
~ f
. ~
:
W
,
.
~ ~
L .
.
~
-
~
,~ ........
~
9
~ ~ ~ ",~: ~z~ ~ :.~ =, ~" ~ . ~ ~ ~ ~
~.~9~
~ ~.~e~g-~e~#~.~,~,~:~-.~.
~-~ e
~~ ~
~:~¢~o~ "~ .
~L~ ~
.
! o J ~ : [ ~~. ~
-
~
e