Breakthrough Cancer Pain: An Observational Study of 1000 European Oncology Patients

Transcription

1 Vol. 46 No. 5 November 2013 Journal of Pain and Symptom Management 619 Original Article Breakthrough Cancer Pain: An Observational Study of 1000 European Oncology Patients Andrew Davies, FRCP, Alison Buchanan, BSc, Giovambattista Zeppetella, FRCP, Josep Porta-Sales, MD, Rudolf Likar, MD, Wolfgang Weismayr, MD, Ondrej Slama, MD, Tarja Korhonen, MD, Marilene Filbet, MD, Philippe Poulain, MD, Kyriaki Mystakidou, MD, Alexandros Ardavanis, MD, Tony O Brien, FRCPI, Pauline Wilkinson, FRCP, Augusto Caraceni, MD, Furio Zucco, MD, Wouter Zuurmond, MD, Steen Andersen, MD, Anette Damkier, MD, Tove Vejlgaard, MSc, Friedemann Nauck, MD, Lukas Radbruch, MD, Karl-Fredrik Sjolund, MD, and Mariann Stenberg, MD Royal Surrey County Hospital (A.Dav.), Guildford, Surrey, United Kingdom; University of Surrey (A.B.), Guildford, Surrey, United Kingdom; St. Clare Hospice (G.Z.), Hastingwood, United Kingdom; Catalan Institute of Oncology (J.P.-S.), Bellvitge Biomedical Research Institute, Barcelona, Spain; Centre for Interdisciplinary Pain Management, Oncology and Palliative Care (R.L.), Klagenfurt, Austria; Mobile Palliative Care Unit (W.W.), Hospice Upper Austria, Salzkammergut, Austria; Masaryk Memorial Cancer Center (O.S.), Brno, Czech Republic; Tampere University Hospital (T.K.), Tampere, Finland; University Hospital Lyon-Sud (M.F.), Lyon, France; Polyclinique de l Ormeau (P.P.), Tarbes, France; University of Athens (K.M.), Athens, Greece; St. Savvas Anticancer Hospital (A.A.), Athens, Greece; Marymount Hospice (T.O.), Cork, Republic of Ireland; Marie Curie Hospice (P.W.), Belfast, United Kingdom; Virgilio Floriani Hospice and Palliative Care Unit (A.C.), National Cancer Institute, Milan, Italy; Department of Anaesthesia, Intensive Care, Palliative Care, Pain Therapy, Hospice and Hospital at Home (F.Z.), Azienda Ospedaliera G. Salvini, Milan, Italy; VU University Medical Center (W.Z.), Amsterdam, The Netherlands; Roskilde Hospital (S.A.), Roskilde, Denmark; Odense University Hospital (A.Dam.), Odense, Denmark; Vejle Hospital (T.V.), Vejle, Denmark; Department of Palliative Medicine (F.N.), University Hospital G ottingen, G ottingen, Germany; University Hospital Bonn (L.R.), Bonn, Germany; Karolinska University Hospital (K.-F.S.), Stockholm, Sweden; and University Hospital Lund (M.S.), Lund, Sweden Abstract Context. Breakthrough pain is common in patients with cancer and is a significant cause of morbidity in this group of patients. Objectives. The aim of this study was to characterize breakthrough pain in a diverse population of cancer patients. Methods. The study involved 1000 cancer patients from 13 European countries. Patients were screened for breakthrough pain using a recommended diagnostic algorithm and then questioned about the characteristics and management of their pain. Address correspondence to: Andrew Davies, FRCP, Royal Surrey County Hospital, Egerton Road, Guildford, Surrey GU2 7XX, United Kingdom. adavies12@nhs.net Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Accepted for publication: December 21, /$ - see front matter

2 620 Davies et al. Vol. 46 No. 5 November 2013 Results. Of the 1000 patients, 44% reported incident pain, 41.5% spontaneous pain, and 14.5% a combination. The median number of episodes was three a day. The median time to peak intensity was 10 minutes, with the median for patients with incident pain being five minutes (P < 0.001). The median duration of untreated episodes was 60 minutes, with the median for patients with incident pain being 45 minutes (P ¼ 0.001). Eight hundred six patients stated that pain stopped them doing something, 66 that it sometimes stopped them doing something, and only 107 that it did not interfere with their activities. Patients with incident pain reported more interference with walking ability and normal work, whereas patients with spontaneous pain reported more interference with mood and sleep. As well, 65.5% of patients could identify an intervention that improved their pain (29.5%, pharmacological; 23%, nonpharmacological; 12%, combination). Regarding medications, 980 patients were receiving an opioid to treat their pain, although only 191 patients were receiving a transmucosal fentanyl product licensed for the treatment of breakthrough pain. Conclusion. Breakthrough cancer pain is an extremely heterogeneous condition. J Pain Symptom Manage 2013;46:619e628. Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Pain, breakthrough pain, cancer, palliative care Introduction Pain is a common symptom in patients with cancer, with an overall prevalence of 53% in studies involving unselected cancer patients. 1 Cancer pain can be classified according to a number of features (e.g., etiology, pathophysiology) and is often classified according to temporal characteristics. Thus, some patients experience an intermittent type of pain (transient pain), although most patients experience a more constant type of pain (background pain). Background pain has been defined as constant or continuous pain of long duration, with the phrase long duration referring to a period of 12 or more hours per day. 2 Breakthrough pain has been defined as a transient exacerbation of pain that occurs either spontaneously, or in relation to a specific predictable or unpredictable trigger, despite relatively stable and adequately controlled background pain. 3 However, there is a lack of consistency in the use of the term breakthrough pain within clinical practice and also within the medical literature. Indeed, the term is widely used to describe any exacerbation of pain in patients with background pain or even intermittent episodes of pain in patients without background pain. Breakthrough pain is common in patients with cancer pain (40e80%) 4 and is a significant cause of morbidity in this group of patients. 5,6 Breakthrough pain is invariably classified according to its relationship to specific events: 1) spontaneous pain (also known as idiopathic pain)dthis type of pain occurs unexpectedly and 2) incident pain (also known as precipitated pain, or, when appropriate, movement-related pain)dthis type of pain is related to specific events and can be subclassified into three categories: volitional, nonvolitional, and procedural. 3 The aim of this study was to characterize breakthrough pain in a diverse population of cancer patients (13 European countries), using a recommended diagnostic algorithm to screen patients for the presence of breakthrough pain. 3 It should be noted that most previously published studies are small in number and involve a single research center. 7,8 Furthermore, most previously published studies do not report the diagnostic criteria used to screen patients for the presence of breakthrough pain, and it appears that at least some of the previously published studies included patients with inadequately controlled background pain (and, therefore, not breakthrough pain).

3 Vol. 46 No. 5 November 2013 European Observational Study of Breakthrough Cancer Pain 621 Methods This was a multicenter study, which involved 1000 patients, 28 specialist palliative care/pain units (see Acknowledgments section), and 13 European countries (Austria, Czech Republic, Denmark, Finland, France, Germany, Greece, Italy, The Netherlands, Republic of Ireland, Spain, Sweden, and United Kingdom). The study was approved by the relevant local research ethics committees and conducted in accordance with the relevant international, European, and national standards for clinical research (e.g., International Conference on Harmonization Good Clinical Practice principles). Patients were given an information sheet in their native language, given the opportunity to ask questions about the study, and then asked to sign a consent form in their native language. Subjects were recruited from patients already receiving treatment in the 28 specialist palliative care/pain units during the period 2008e2011. Subjects were recruited in a consecutive manner from both the inpatient and outpatient populations. The inclusion criteria for the study were as follows: 1) age older than 18 years, 2) cancer diagnosis, 3) cancerrelated pain, 4) cancer-related breakthrough pain, and 5) regular use of an opioid for moderate-to-severe pain (a so-called strong opioid). Fig. 1 shows the diagnostic algorithm used to identify patients with breakthrough pain. 3 The exclusion criteria for the study were cognitive impairment and limited comprehension of the native language. The questionnaire was specifically developed for the purposes of the present study. The first Fig. 1. Diagnostic algorithm used to identify patients with breakthrough pain (adapted from Davies et al. 3 ).

4 622 Davies et al. Vol. 46 No. 5 November 2013 section consisted of the screening questions to identify patients with breakthrough pain (Fig. 1); patients who failed the screening questions did not complete the remainder of the questionnaire. The second section asked questions about the characteristics of the breakthrough pain, including interference with various aspects of daily living (i.e., general activity, mood, walking ability, normal work [includes both work outside the home and housework], relations with other people, sleep, and enjoyment of life). 9 The interference with daily activities was assessed using a 0e10 numerical rating scale, where 0 represented does not interfere and 10 represented completely interferes. The third section asked questions about the current treatment of the breakthrough pain. Patients also were asked about their adherence with rescue medication for breakthrough pain (and reasons for nonadherence). Statistical Analysis The data were translated into English (where necessary), then were entered into a secure data management system (at the Royal Marsden Hospital), and subsequently were analyzed using IBM SPSS Statistics, version 18 (IBM Corporation, Armonk, NY). Descriptive statistics were used to describe the main study results (e.g., median value with lower and upper value ranges). Missing data were reported as such, rather than using any specific statistical techniques. Comparisons were made between patients with incident-type breakthrough pain and patients with spontaneous-type breakthrough pain. Chi-squared tests were used to compare categorical variables, and Mann-Whitney U tests were used to compare continuous variables. The Kruskal-Wallis test was used to compare the frequency of breakthrough pain episodes in patients with differing Eastern Cooperative Oncology Group (ECOG) performance status, and then pairwise comparisons were undertaken to further elucidate the relationship between these variables (with the Bonferroni method used to correct for multiple testing). Results One thousand patients were recruited to the study, with 60 patients being recruited from the Czech Republic and Finland and 80 patients being recruited from each of the other countries. The median age of the subjects was 62 years (range 23e93 years), and there was a slight preponderance of male subjects (51% of total). The data on ECOG performance status and cancer diagnosis are summarized in Table 1. Four hundred forty patients (44%) reported incident-type breakthrough pain (i.e., the episodes were related to an identifiable precipitant), 415 (41.5%) reported spontaneous-type breakthrough pain (i.e., the episodes were not related to an identifiable precipitant), 143 (14.5%) reported a combination of incident- and spontaneous-type breakthrough pain, and there was no information on two patients. The incident-type pains were volitional in 324 patients (i.e., brought on by a voluntary act such as weightbearing/walking), nonvolitional in 67 patients (i.e., brought on by an involuntary act such as coughing/vomiting), a combination of volitional and nonvolitional in 35 patients, procedural in five patients, and not specified in nine patients. The median number of episodes of breakthrough pain was three per day (range, one per month to 24 per day), and there was no difference between patients with incidenttype pain and those with spontaneous-type pain. However, the ECOG performance status did influence the frequency of breakthrough pain episodes (Kruskal-Wallis test; P < 0.001) (Table 2). Nine hundred thirty-six patients were able to answer the question about the time to peak intensity; the median for the whole group was 10 minutes (range, less than one minute to 240 minutes), with the median for patients with incident-type pain being five minutes (range, less than one minute to 240 minutes) and the median for patients with spontaneous-type pain being 10 minutes (range, less than one minute to 240 minutes) (Fig. 2). The difference between patients with incident-type pain and those with spontaneous-type pain was statistically significant (Mann-Whitney U test; P < 0.001). In contrast, only 505 patients were able to answer the question about the duration of untreated episodes. The median for the whole group was 60 minutes (range, less than one minute to 360 minutes), with the median for patients with incident-type pain being 45 minutes

5 Vol. 46 No. 5 November 2013 European Observational Study of Breakthrough Cancer Pain 623 Table 1 Characteristics of Subjects Characteristic All Patients (N ¼ 1000) a Patients With Incident Pain (n ¼ 440) Patients With Spontaneous Pain (n ¼ 415) Patients With Mixed Pain (n ¼ 143) ECOG Performance Status ECOG ECOG ECOG ECOG ECOG Not stated Cancer diagnosis Breast Gastrointestinal Gynecological Hematological Head and neck Lung Melanoma Neurological Sarcoma Urological Unknown Not stated Background analgesia Opioid for moderate-to-severe pain Opioid for mild-to-moderate pain Opioid not stated Nonopioid b Adjuvant analgesic Breakthrough analgesia Opioid for moderate-to-severe pain Opioid for mild-to-moderate pain Opioid not stated No opioid Nonopioid b Adjuvant analgesic ECOG ¼ Eastern Cooperative Oncology Group. a Includes two patients with indeterminate type of breakthrough pain. b Nonopioid ¼ paracetamol (acetaminophen) and nonsteroidal anti-inflammatory drugs. (range, less than one minute to 360 minutes) and the median for patients with spontaneoustype pain being 60 minutes (range, less than one minute to 360 minutes) (Fig. 3). The difference between patients with incident-type pain and those with spontaneous-type pain was again statistically significant (Mann-Whitney U test; P ¼ 0.001). The intensity of the breakthrough pain was rated as mild in 36 cases, moderate in 337 cases, and severe in 618 cases (with no specific information for nine patients). There was no difference between patients with incident-type pain and those with spontaneoustype pain. Unsurprisingly, 806 patients stated that the pain stopped them doing something, 66 that the pain sometimes stopped them doing something, and only 107 said that the pain did not interfere with their activities of daily living. Patients with incident-type pain more commonly reported that the pain stopped them doing something than patients with spontaneous-type pain (Chi-squared test; P ¼ 0.018). Fig. 4 and Table 3 show data on the degree of interference with various aspects of daily living. Patients with incident-type pain reported more interference with walking ability and normal work, whereas patients with spontaneous-type pain reported more interference with mood and sleep. Most (65.5%) patients could identify an intervention that usually improved the breakthrough pain, although another 10.5% of patients could identify an intervention that sometimes improved the breakthrough pain. In other words, 23.5% of patients had found nothing that relieved their breakthrough pain. The interventions that were usually successful were pharmacological in 294 cases, nonpharmacological in 232 cases, a

6 624 Davies et al. Vol. 46 No. 5 November 2013 Table 2 ECOG Performance Status and Frequency of Breakthrough Pain Number of Breakthrough Cancer Pain Episodes ECOG Performance Status Mean (95% CI) Pairwise Comparison (1.60e2.59) d (2.62e3.24) ECOG 1 z ECOG 0 (P ¼ 0.70) (3.15e3.73) ECOG 2 > ECOG 0 (P ¼ 0.02) ECOG 2 z ECOG 1 (P ¼ 0.09) (3.65e4.35) ECOG 3 > ECOG 0 (P ¼ 0.001) ECOG 3 > ECOG 1 (P < 0.001) ECOG 3 z ECOG 2 (P ¼ 0.40) (2.99e4.01) ECOG 4 > ECOG 0 (P ¼ 0.03) ECOG 4 z ECOG 1 (P ¼ 0.38) ECOG 4 z ECOG 2 (P ¼ 1.00) ECOG 4 z ECOG 3 (P ¼ 1.00) ECOG ¼ Eastern Cooperative Oncology Group; 95% CI ¼ 95% confidence interval. combination of pharmacological and nonpharmacological in 120 cases, and unspecified in 10 cases. Patients with incident-type pain were more likely to report a relieving factor (Chi-squared test; P < 0.001) and also to report the success of nonpharmacological interventions (either alone or in combination with medication) (Chi-squared test; P < 0.001). Table 1 shows the medication prescribed to the patients to manage their breakthrough pain (and also their background pain), and Table 4 shows the nonpharmacological interventions used by the patients to manage their breakthrough pain. Nine hundred eighty patients were receiving an opioid to treat their breakthrough pain, although only 191 patients were receiving one of the transmucosal fentanyl products licensed for the management of breakthrough pain. The complementary therapies used were acupuncture (n ¼ 4), aromatherapy (n ¼ 9), healing (n ¼ 1), homeopathy (n ¼ 3), and Reiki (n ¼ 1). The other nonpharmacological interventions used by more than one individual were hypnosis (n ¼ 2), prayer (n ¼ 4), drinking tea/coffee (n ¼ 9), and drinking alcohol (n ¼ 2). Interestingly, only 531 (53%) patients took their rescue medication every time they experienced breakthrough pain, although the reasons for nonadherence were many and varied. Two hundred seven patients stated that one of the reasons was that the pain was not always severe (with no difference in response between patients with incident-type or spontaneous-type pain). Similarly, 174 patients Fig. 2. Time to peak intensity. BTcP ¼ breakthrough cancer pain.

7 Vol. 46 No. 5 November 2013 European Observational Study of Breakthrough Cancer Pain 625 Fig. 3. Duration of untreated episodes. BTcP ¼ breakthrough cancer pain. stated that one of the reasons was that the pain was not always long-lasting (with again no difference in response between patients with incident-type or spontaneous-type pain). Discussion This is a unique study in terms of magnitude (1000 patients), setting (28 centers; 13 countries), and methodology used (diagnostic algorithm). It is clear from the data that breakthrough pain is a heterogeneous condition and that episodes vary between individuals and also within individuals (at the same time). Equally, it is clear from the data that breakthrough pain has a significant negative effect on quality of life and that this is related to a direct effect (suffering) and an indirect effect (interference with activities of daily living). The classic depiction of breakthrough pain is acute in onset, short in duration, and moderate to severe in intensity. 3 The median time to peak intensity was 10 minutes in this study, which is somewhat longer than that reported in previous studies (i.e., one to three minutes). 5,10 It should be noted that many patients commented that the pain became intense shortly after the start of the episode, even if the pain did not peak in intensity until 10 minutes after the start of the episode. The median duration of untreated episodes was 60 minutes in this study, which is in keeping with some previous studies, 10,11 although somewhat Fig. 4. Box and whisker plot showing lower quartile, median, upper quartile, and range of values for interference with various aspects of daily living.

8 626 Davies et al. Vol. 46 No. 5 November 2013 Interference Scale Table 3 Interference Caused by Breakthrough Pain All Patients Patients With Incident Pain Patients With Spontaneous Pain Median (Range) Median (Range) Median (Range) P-value (Difference for Patients With Incident/ Spontaneous Pain) General activity 7 (0e10) 7 (0e10) 7 (0e10) Mood 7 (0e10) 6 (0e10) 7 (0e10) Walking ability 7 (0e10) 8 (0e10) 6 (0e10) <0.001 Normal work (includes both work 8(0e10) 9 (0e10) 8 (0e10) outside the home and housework) Relations with other people 5 (0e10) 5 (0e10) 5 (0e10) Sleep 5 (0e10) 5 (0e10) 6 (0e10) <0.001 Enjoyment of life 7 (0e10) 7 (0e10) 7 (0e10) longer than that reported in other studies. 5,6,12e14 Interestingly, only w50% of patients were able to answer this question (in contrast to w94% of patients who were able to answer the question about the time to peak intensity). The data on pain intensity in this study were entirely consistent with the data on pain intensity in previous studies. 11,13,14 As discussed, breakthrough pain can be classified into either spontaneous-type pain (i.e., no obvious precipitant) or incident-type pain (i.e., an obvious precipitant). Patients with incident-type pain had a shorter time to peak intensity (5 minutes vs. 10 minutes) and a shorter duration of untreated episodes (45 minutes vs. 60 minutes). The published literature contains little analogous data, but the findings are consistent with those reported by Spanish researchers. 14 Patients with incidenttype pain were more likely to report interference with activities of daily living and particularly with walking ability and normal work. It should be noted that many of the patients with incident-type pain had bone metastases and that the incident-type pain was precipitated by weightbearing/movement. In contrast, patients with spontaneous-type pain were more likely to report interference with sleep and mood. The published literature contains no analogous data, although it has been previously suggested that spontaneous-type pain may have a greater impact on function than incident pain (because of its unpredictable nature). 15 Sixty-five percent of patients had identified an intervention that improved their breakthrough pain; the intervention was pharmacological in 29% of subjects, nonpharmacological in 23% of subjects, and a combination of pharmacological and nonpharmacological interventions in 12% of subjects. The apparent ineffectiveness of breakthrough ( rescue ) medication has been reported in previous Table 4 Nonpharmacological Interventions Used by Patients to Manage Breakthrough Pain Intervention a No. of Patients (N ¼ 1000) None 618 Not stated 2 Rest/sleep 103 Movement/exercise 42 Change of position 94 Physical support 13 Heat 107 Cold 22 Rubbing/massage 36 Transcutaneous electrical nerve stimulation 14 Relaxation/visualization 24 Distraction 22 Miscellaneous complementary therapies (see text) 18 Miscellaneous other interventions (see text) 22 a Patient could be using more than one intervention.

9 Vol. 46 No. 5 November 2013 European Observational Study of Breakthrough Cancer Pain 627 studies 5,6,12 and is undoubtedly related to a combination of different factors (i.e., pain related and analgesic related). For example, some breakthrough pains are so transient in nature that they are unlikely to respond to any type of pharmacological intervention. Similarly, some breakthrough pains are relatively unresponsive to opioid analgesics ( opioid poorly responsive pain ), which are the cornerstones of the management of breakthrough pain episodes. 3 The study results reflect the experiences of a large cohort of European oncology patients who were receiving care from a specialist palliative care or a pain team. However, the study population may not be entirely representative of the wider oncology population (by nature of the fact that they had been referred to a specialist palliative care team or a pain team). Moreover, the study population included few patients with an ECOG performance status of 0 (3.5% of total) or an ECOG performance status of 4 (7.5% of total). Conclusion Breakthrough pain is a heterogeneous condition, and the episodes vary both between individuals and within individuals. Nevertheless, breakthrough pain often has a significant negative effect on quality of life, and this is related to a direct effect (suffering) and an indirect effect (interference with activities of daily living). Disclosures and Acknowledgments This study was funded by Dr. Davies s Palliative Care Research Fund at the Royal Marsden Hospital. The study was supported by an unrestricted educational grant from Nycomed. Nycomed was not involved in the design of the study, collection of the data, analysis of the data, interpretation of the data, writing of the paper, or decision to submit the paper for publication. None of the authors have a conflict of interest relevant to this study or this publication. The authors would like to acknowledge the assistance of all the staff at the research institutions: Royal Marsden Hospital, Sutton, U.K.; St. Clare Hospice, Hastingwood, U.K.; ZISOP, Klagenfurt, Austria; Dachverband Hospiz Osterreich, V ocklabruck, Austria; Masaryk Oncology Institute, Brno, Czech Republic; University Hospital Hradec Kralove, Hradec Kralove, Czech Republic; Tampere University Hospital, Tampere, Finland; Helsinki University Hospital, Helsinki, Finland; University Hospital Lyon-Sud, Lyon, France; Polyclinique de l Ormeau, Tarbes, France; Pain Relief and Palliative Care Unit of the University of Athens, Athens, Greece; Aghios Savvas Anticancer Hospital, Athens, Greece; Marie Curie Hospice, Belfast, Northern Ireland; Marymount Hospice, Cork, Republic of Ireland; Our Lady s Hospice, Dublin, Republic of Ireland; National Cancer Institute, Milan, Italy; Azienda Ospedaliera G. Salvini, Milan, Italy; VU University Medical Center, Amsterdam, The Netherlands; Renier de Graaf Groep, Delft, The Netherlands; Catalan Institute of Oncology, Barcelona, Spain; Hospital General Universitario Gregorio Maranon, Madrid, Spain; Odense University Hospital, Odense, Denmark; Roskilde Hospital, Roskilde, Denmark; Sankt Maria Hospice, Vejle, Denmark; University Hospital Aachen, Aachen, Germany; University Hospital G ottingen, G ottingen, Germany; Karolinska University Hospital, Stockholm, Sweden; and University Hospital Lund, Lund, Sweden. The authors specifically acknowledge the assistance of Dr. Jana Hrubesova, Dr. Tina Saarto, Dr. Michel Wagemans, Dr. Juan Nunez Olarte, and Kabir Mohammed Batsari (medical statistician). References 1. van den Beuken-van Everdingen MH, de Rijke JM, Kessels AG, et al. Prevalence of pain in patients with cancer: a systematic review of the past 40 years. Ann Oncol 2007;18:1437e Ferrell BR, Juarez G, Borneman T. Use of routine and breakthrough analgesia in home care. Oncol Nurs Forum 1999;26:1655e Davies AN, Dickman A, Reid C, Stevens AM, Zeppetella G. The management of cancer-related breakthrough pain: recommendations of a task group of the Science Committee of the Association for Palliative Medicine of Great Britain and Ireland. E J Pain 2009;13:331e Mercadante S, Radbruch L, Caraceni A, et al. Episodic (breakthrough) pain: consensus conference of an expert working group of the European Association for Palliative Care. Cancer 2002;94: 832e839.

10 628 Davies et al. Vol. 46 No. 5 November Portenoy RK, Payne D, Jacobsen P. Breakthrough pain: characteristics and impact in patients with cancer pain. Pain 1999;81: 129e Hwang SS, Chang VT, Kasimis B. Cancer breakthrough pain characteristics and responses to treatment at a VA medical center. Pain 2003;101: 55e Zeppetella G, Ribeiro MD. Pharmacotherapy of cancer-related episodic pain. Expert Opin Pharmacother 2003;4:493e Davies A. Cancer-related breakthrough pain, 2nd ed. Oxford: Oxford University Press, Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Inventory. Ann Acad Med Singapore 1994;23:129e Portenoy RK, Bruns D, Shoemaker B, Shoemaker SA. Breakthrough pain in communitydwelling patients with cancer pain and non-cancer pain, part 1: prevalence and characteristics. J Opioid Manag 2010;6:97e Fine PG, Busch MA. Characterization of breakthrough pain by hospice patients and their caregivers. J Pain Symptom Manage 1998;16:179e Portenoy RK, Hagen NA. Breakthrough pain: definition, prevalence and characteristics. Pain 1990;41:273e Zeppetella G, O Doherty CA, Collins S. Prevalence and characteristics of breakthrough pain in cancer patients admitted to a hospice. J Pain Symptom Manage 2000;20:87e Gomez-Batiste X, Madrid F, Moreno F, et al. Breakthrough cancer pain: prevalence and characteristics in patients in Catalonia, Spain. J Pain Symptom Manage 2002;24:45e Portenoy RK. Treatment of temporal variations in chronic cancer pain. Semin Oncol 1997;5(S16): 7e12.