Vol. 18 No. 1 July 1999 Journl of Pin nd Symptom Mngement 9 Originl Article Continuous Subcutneous Infusion of Morphine vs. Hydromorphone: A Controlled Tril Mry G. Miller, MB, MRCP (Irelnd), Noel McCrthy, BA, MB, MRCP (Irelnd), Cirán A. O Boyle, BSc, PhD, nd Michel Kerney, MB, FRCP (Irelnd) Sir Michel Sobell House (M. G. M.), Churchill Hospitl, Oxford, United Kingdom; Our Ldy s Hospice (M.G.M., M.K.), Dublin, Irelnd; Smittskyddsinstitutet (N.M.), Stockholm, Sweden; Deprtment of Psychology (C.A.O.), Royl College of Surgeons, Dublin, Irelnd Abstrct Seventy-four ptients were included in double-blind, rndomized, controlled tril compring the nlgesic efficcy nd dverse effects of hydromorphone nd morphine delivered by continuous subcutneous infusion. Ptients completed the Memoril Pin Assessment Crd nd checklist of opioid-relted dverse effects immeditely before commencing subcutneous infusion nd 24, 48, nd 72 hours lter. An ssessment tool ws developed for the 60 ptients who were too ill to complete their own questionnire. The tool demonstrted excellent inter-rter relibility. Thirty-four percent of ptients in the hydromorphone group nd 27% of those in the morphine group died before completion of the study (P 0.66). The hydromorphone group required more nlgesi for brekthrough pin in the first 24 hours of the study (P 0.03) nd hd greter improvement in the behvior of frowning on movement nd the comfort visul nlogue scle (P 0.08) over the course of the study. Adverse effects were rre nd similr in both groups. This study found hydromorphone to be t lest s effective s morphine when delivered by continuous subcutneous infusion. J Pin Symptom Mnge 1999;18: 9 16. U.S. Cncer Pin Relief Committee, 1999. Key Words Pllitive, hospice, hydromorphone, morphine, syringe driver, subcutneous infusion, proxy ssessment, pin. Introduction Hydromorphone is strong opioid, semisynthetic derivtive of morphine tht ws introduced s n nlgesic during the second decde of this century. 1,2 Hydromorphone is used s n lterntive to morphine in countries where dimorphine is not licensed for medicl Address reprint requests to: Mry G. Miller, MB, MRCP, Sir Michel Sobell House, Churchill Hospitl, Oxford, OX3 7LJ, United Kingdom. Accepted for publiction: September 23, 1998. use. Hydromorphone is both more soluble 3 nd more potent thn morphine, 4 llowing equinlgesic doses to be infused subcutneously in smller volumes of solution. A smller volume for infusion requires less frequent chnge of syringe driver sites nd mediction syringes, n dvntge for both ptient nd stff. Morphine nd hydromorphone re used interchngebly in the syringe driver. At equinlgesic doses the dverse effect profiles re expected to be similr, 5 s both drugs hve similr chemicl structure 6 8 nd similr phrmcodynmic nd phrmcokinetic profiles. 9 U.S. Cncer Pin Relief Committee, 1999 0885-3924/99/$ see front mtter Published by Elsevier, New York, New York PII S0885-3924(99)00036-6
10 Miller et l. Vol. 18 No. 1 July 1999 To dte there is no empiricl evidence to support this prctice. The im of this study ws to compre the nlgesic efficcy nd dverse effect profile of hydromorphone nd morphine given by subcutneous infusion. Methods Design nd Ptient Selection A rndomized, double-blind protocol ws used to compre hydromorphone with morphine, the recommended tretment for severe cncer pin. 10 Ptients were considered for entry t the time of their dmission to the hospice inptient unit during the 12-month study period. The objectives of the study were discussed with ptients nd their fmilies. Informed consent ws sought for entry to the study in the event of the ptient being unble to tke orl medictions nd requiring continuous subcutneous infusion of strong opioid during their dmission. Ptients who hd subcutneous infusion t the time of dmission, who were receiving opioids by nother prenterl route (e.g., epidurl infusion), who hd history of intolernce to either drug, or who hd cognitive impirment nd were therefore unble to give informed consent were excluded. A ptient could lso be excluded t the discretion of the stff if the ptient or his or her fmily were very distressed. The study protocol nd consent form were reviewed by externl scientific dvisors nd pproved by St. Vincent s Hospitl ethics committee. Tretment Protocol Ptients who hd given informed consent were rndomized to receive either hydromorphone or morphine, if decision ws mde by medicl stff tht continuous subcutneous infusion of n nlgesic ws required. The orl to subcutneous rtio for morphine ws regrded s 2:1 11 nd the rtio of subcutneous morphine to hydromorphone s 5:1. 1,4 Morphine sulfte ws given orlly for brekthrough pin. If unble to ingest or bsorb it, the ptient ws given subcutneous morphine or hydromorphone ccording to the ctegory of rndomiztion. Hydromorphone ws not vilble for orl use in Irelnd t this time. The nlgesic dose used to mnge brekthrough pin ws one-sixth of the totl dily dose. Assessment of Pin nd Adverse Effects Pin nd dverse effects were ssessed immeditely before commencing the subcutneous infusion, nd 24, 48, nd 72 hours lter. On rre occsions when n infusion ws commenced fter 10 pm, the initil ssessment took plce t tht time but the subsequent ssessment ws deferred until fter 8 m in order not to disturb the ptient s sleep. Ptients who were well enough to do so completed subjective ssessment of pin using the Memoril Pin Assessment Crd 12 (Appendix 1). This mesure is vlid for use in cncer pin, nd is brief nd simple. It is, therefore, useful in this group of very ill ptients. 13 The presence or bsence of nuse, constiption, myoclonus, generlized pruritus, tinnitus, nd ny erythem, tenderness, or pruritus t the syringe driver site were ssessed by the ptient (Appendix 1) t ech of the study times. The study popultion ws very ill, nd it ws nticipted tht the mjority of ptients would be unble to complete subjective ssessment of pin nd dverse effects. Assessment of pin by proxy is possible by two mens, nlgesic consumption nd stff or fmily member ssessment of pin. 14 A tool ws developed bsed on stff ssessment prctice in the hospice (Appendix 2). The tool incorported the observtion of the presence or bsence of five ptient behviors nd the completion of three visul nlogue scles to ssess the level of the ptient s pin. The presence or bsence of retching or vomiting, myoclonus, nd erythem or discomfort of the syringe driver site ws lso noted. Assessment t ech of the nominted intervls ws Tble 1 Resons for Exclusion of Ptients (n = 57) from the Study Reson Number A. Excluded t dmission Subcutneous infusion in plce on dmission 13 Emotionl distress 9 Consent refused 3 Epidurl infusion in plce 2 B. Excluded t rndomiztion Filure to consider for study 24 No reson recorded 5 Nurse resercher unvilble 2 Two resons given for one ptient.
Vol. 18 No. 1 July 1999 Subcutneous Hydromorphone vs. Morphine 11 crried out by the nurse cring for the ptient t tht time. This nurse ws not responsible for dministrtion of mediction, nd tretment remined blinded. The inter-rter relibility of the tool ws tested by gthering dt from two simultneous but independent ssessments of five ptients, t ech of the four study times. A reserch nurse recorded demogrphic dt nd pin history from ech ptient s notes nd gthered ll ssessment nd nlgesic consumption dt. Unexpected nd dverse events were lso recorded. Results Smple Size In totl, 133 ptients required mediction by subcutneous infusion during the study period. Fifty-seven ptients were excluded prior to rndomiztion. The resons for their exclusion re listed in Tble 1. Forty-one ptients were rndomized to receive morphine nd 36 to receive hydromorphone. Three ptients, rndomized to the hydromorphone group, received morphine s brekthrough nlgesi in Tble 2 Ptient Chrcteristics Hydromorphone (n = 33) Morphine (n = 41) Totl (n = 74) Age (yers) Medin 67 71 69 Rnge 35 85 34 88 34 88 Sex Mle 11 (33%) 22 (54%) 33 (45%) Femle 22 (67%) 19 (46%) 41 (55%) Dignosis Brest crcinom 11 (33%) 7 (17%) 18 (24%) Lung crcinom 7 (21%) 7 (17%) 14 (19%) Colorectl crcinom 5 (16%) 5 (12%) 10 (13%) Prosttic crcinom 0 (0%) 8 (20%) 8 (11%) Nsophryngel crcinom 2 (6%) 3 (7%) 5 (7%) Pncretic crcinom 2 (6%) 2 (5%) 4 (5%) Renl crcinom 1 (3%) 2 (5%) 3 (4%) Esophgel crcinom 0 (0%) 3 (7%) 3 (4%) AIDS 1 (3%) 1 (3%) 2 (3%) Others 4 (12%) 3 (7%) 7 (10%) Indiction for syringe driver Too wek to tke orl medicine 24 30 54 Nuse nd vomiting 9 12 21 Too mny tblets 1 3 4 Others 1 3 4 Suspected bowel obstruction 1 2 3 Unspecified 0 1 1 Number of pin sites b 1 11 (33%) 18 (44%) 29 (39%) 2 17 (52%) 18 (44%) 35 (47%) 3 5 (15%) 5 (12%) 10 (14%) Cuse of pin Bony metstses 13 17 31 Debility 13 17 31 Locl extension of cncer 9 16 25 Soft tissue metstses 10 6 16 Constiption 7 2 9 Neurl compression/invsion 3 2 5 Incresed intrcrnil pressure 2 2 4 Others 5 5 10 Ptient tking strong opioid c No 8 6 14 Yes 25 35 60 Medin dose (Rnge of dose) 40 (0 780) 40 (0 350) 40 (0 780) Others: gitted 1; petit-ml seizures 1; risk of hemorrhge 1; pin control difficult to chieve 1. 13 ptients hd two resons for commencing subcutneous infusion. b No ptient reported more thn three sites of pin. c Doses meured in milligrms nd converted to the equivlent orl morphine dose, tken by the ptient in the 24 hours before rndomiztion.
12 Miller et l. Vol. 18 No. 1 July 1999 Brekthrough nlgesics (prn) Tble 3 Anlgesic Consumption: The Use of Brekthrough Anlgesi From 0 to 24 hours Proportion of ptients receiving prn doses Men number of prn doses per ptient Proportion of ptients receiving prn doses From 24 to 72 hours Men number of prn doses per ptient Hydromorphone 14/29 (48%) 0.9 7/23 (30%) 0.8 Morphine 8/37 (22%) 0.5 11/30 (37%) 0.5 Comprison b RR 2.2 (1.1 4.6) P = 0.03 RR 0.8 (0.4 1.8) P = 0.5 Denomintor of the proportion represents the number of ptients live t tht point of the study. b RR (95% confidence intervls) = reltive risk of n event for those ptients tking hydromorphone compred to those tking morphine. error nd were withdrwn from the study. Dt ws vilble on 74 ptients for nlysis. Ptient Chrcteristics Ptient chrcteristics re described in Tble 2. The medin ge ws 69 yers, nd 55% were women. Seventy-two ptients hd dignosis of cncer nd two dignosis of AIDS. The most common reson for commencing subcutneous infusion ws tht the ptient ws too wek to tke orl medictions (73%). Forty-nine percent of the ptients hd 2 sites of pin. Metstses to bone, debility, nd locl invsion by the cncer were the most common cuses of pin. Sixty ptients (82%) hd tken strong opioid before entry into the study. The medin dose for both groups ws the equivlent of 40 mg of orl morphine in the 24 hours before rndomiztion. Twenty-six ptients were tking djuvnt nlgesics. Attrition Rte Deth ws the only reson for premture exit from the study. Thirty percent of ptients in the hydromorphone group nd 27% in the morphine group died before completion of the study (P 0.66). Eight ptients died before the second ssessment t 24 hours, 7 before 48 hours, nd 6 before the finl ssessment t 72 hours. All four ssessments were completed for or by the remining 53 people (72%). Tool Designed for Assessment by Proxy of Ptient Pin There ws complete correltion between the ssessments of behvior, i.e., rigidity, moning, nd frowning. There ws excellent correltion between the visul nlogue scores of restfulness (correltion coefficient r 0.98), comfort (r 0.95), nd pin (r 0.94). Anlgesic Efficcy s Mesured by Additionl Drug Use During the Study Ptients in the hydromorphone group were twice s likely to need brekthrough nlgesi s those in the morphine group during the initil 24 hours of the study (P 0.03). The mount of brekthrough nlgesi required by both groups ws comprble for the reminder of the study period (Tble 3). Forty-eight percent of ptients in the hydromorphone group nd 35% in the morphine group required n increse in the dose deliv- Anlgesic in syringe driver (SD) Tble 4 Anlgesic Consumption: Chnges in Anlgesi Delivered by Continuous Subcutneous Infusion From 0 to 24 hours From 24 to 72 hours From 0 to 72 hours Proportion of ptients with increse in SD Men increse in nlgesic infused (%) Proportion of ptients with increse in SD Men increse in nlgesic infused (%) Proportion of ptients with increse in SD Men increse in nlgesic infused (%) Hydromorphone 14/29 (48%) 14 10/23 (43%) 10 12/23 (52%) 21 Morphine 13/37 (35%) 17 14/30 (47%) 29 19/30 (63%) 39 Comprison b RR 1.2 (0.6 2.2) P = 0.6 RR 0.9 (0.5 1.7) P = 0.3 RR 0.8 (0.5 1.3) P = 0.3 Denomintor of the proportion represents the number of ptients live t tht point of the study. b RR (95% confidence intervls) = reltive risk of n event for those ptients tking hydromorphone compred to those tking morphine.
Vol. 18 No. 1 July 1999 Subcutneous Hydromorphone vs. Morphine 13 ered by subcutneous infusion t 24 hours (reltive risk [RR] 1.2, Tble 4). The men percentge increse per ptient ws, however, lower in the hydromorphone group (14%) thn in the morphine group (17%) (P 0.6). Over the course of the study, both groups needed n increse in the nlgesi delivered by continuous subcutneous infusion (52% of ptients in the hydromorphone group nd 63% of those in the morphine group) (RR 0.8). The increse in the morphine group ws twice tht of the hydromorphone group (P 0.3). Anlgesic Efficcy s Mesured by Chnges in Pin Assessments Only 8 ptients were ble to complete the Memoril Pin Assessment Crd t ll four designted ssessment times. There ws no significnt difference observed between the hydromorphone nd morphine groups, but this finding is meningless given the smll number of ptients. The chnge in pin over the first 24 hours of the study, s ssessed by proxy, reveled no difference between the group receiving hydromorphone nd tht receiving morphine (Tble 5). Over the course of the 72 hours of the study, hydromorphone demonstrted trend to improve ll pin vribles mesured when compred with morphine. In prticulr, frowning on movement nd the comfort level on visul nlogue score were improved but none of the vribles were sttisticlly significnt t level of P 0.05. Adverse Effects The number of dverse events noted ws smll nd there were no significnt differences between the groups. Discussion The study popultion is representtive of terminlly ill pllitive cre popultion, 28% of ptients dying within the 72-hour period of the study. Sixty-seven percent of the hydromorphone group were women, compred with 47% of the morphine group. This sex distribution is reflected in the ptient dignoses, brest cncer being more common in the hydromorphone group nd prostte cncer in the morphine group. Other differences between the groups were less mrked (Tble 1). None of the differences between the groups were sttisticlly significnt t level of P 0.05. The finding tht ptients in the hydromorphone group were more likely to require nlgesi for brekthrough pin during the first 24 hours of the study (RR 2.2, Tble 3) is interesting. Almost hlf (48%) of ptients in the hydromorphone group required n increse in the continuous subcutneous infusion of nlgesic t 24 hours, compred with 35% in the morphine group, lthough this trend did not ttin sttisticl significnce (RR 1.2, Tble 4). This finding my simply reflect the biologicl phenomenon of wide vrition in potency rtios described in the literture. 15,16 The observtion suggests tht the hydromorphone: morphine potency rtio of 1:5 ws too low for Tble 5 Comprison of the Effect of Hydromorphone nd Morphine on Pin Mesured by Proxy Hydromorphone (n = 23) 0 to 24 hours Morphine Hydromorphone (n = 31) P Vlue c (n = 17) 0 to 72 hours Morphine (n = 25) P Vlue c Rigid on turning 0 0.1 0.34 k 0.18 0 0.32 k Mon t rest 0.04 0.06 0.74 k 0.12 0.12 0.18 k Mon on movement 0 0.13 0.08 k 0.18 0.08 0.18 k Frown t rest 0.13 0 0.11 k 0.29 0 0.14 k Frown on movement 0.09 0 0.25 k 0.29 0 0.08 k Pin on VAS b 1.3 1.1 0.79 3.3 1.9 0.11 Comfort level on VAS b 1.3 1.3 0.97 3.5 1.8 0.07 k Restfulness on VAS b 1.4 1.5 0.87 2.7 2.6 0.94 Vlues represent the difference between ssessment t bseline nd lter. Negtive vlues indicte decrese in the observed behvior, 0 no chnge, nd positive vlues indicte n increse. The rnge of the scle is from 1 to 1. b Vlues represent the difference between ssessment t bseline nd lter. Negtive vlues indicte decrese in the observed behvior, 0 no chnge, nd positive vlues indicte n increse. The rnge of the scle is from 20 to 20. VAS = Visul nlogue scle. c P vlues clculted by ANOVA except when vrinces between the two groups were dissimilr or when the smple dt ws not normlly distributed. In these cses Kruskl-Wllis test ws used (vlues mrked k ).
14 Miller et l. Vol. 18 No. 1 July 1999 lmost one-hlf of the ptients. Potency rtios hve been mostly determined from single-dose studies in postopertive ptients using low-dose opioids. 16 Recent reserch suggests tht empiricl evidence supporting these rtios in pllitive cre setting presents crude nd, on occsion, inccurte picture. 15 Lwlor et l. 16 in their retrospective study clculted potency rtio of 4.3:1 for morphine to hydromorphone. However, tht study excluded ptients who were converted from orl to prenterl nlgesi, s ws the cse in this study. A previous study 15 included these ptients but unfortuntely did not do subgroup nlysis s the numbers were smll. Further prospective reserch would be helpful in estblishing potency rtios of nlgesics in terminlly ill ptients. Frequent review of the ptient, stndrdiztion of ssessment, nd n pprecition of biologicl vrition re importnt fctors in ensuring tht good pin control is chieved rpidly. Over the course of the 72 hours of the study, hydromorphone demonstrted trend to improve ll pin vribles mesured when compred with morphine for those ptients whose pin ws ssessed by stff (P 0.05 for ll vribles). Repliction of this finding is importnt, s it my be influenced by the smll number of ptients in this study. Eighty-five percent of ptients in this study were too ill to complete the Memoril Pin Assessment Crd, nondemnding instrument which cn be completed in less thn 1 minute by ptients fmilir with its use. 12 Assessment by proxy, either ptients or stff, is not idel, 14 but necessry to monitor tretment nd prt of routine clinicl prctice when the ptient is not competent or is too ill to report pin nd the effect of dose djustments. No suitble tools were published in the literture for this specific popultion. A tool ws developed bsed on clinicl prctice in the unit, ssessment of the ptient to determine the cuse of ny observed discomfort, nd intervention to try to correct the cuse. The tool developed for proxy ssessment of the ptient s pin by the nursing stff demonstrted excellent interrter relibility. Questions of vlidity remin more difficult to investigte. Behvior mesured my reflect other cuses of distress, psychologicl s well s physicl, in ddition to the presence of pin. However, this tool reflects nd stndrdizes the clinicl prctice of the unit. Stndrdiztion of pin ssessment for ptients unble to complete self-ssessment is essentil. Conclusion This study demonstrted tht hydromorphone is t lest s effective n nlgesic s morphine when delivered by continuous subcutneous infusion. Frequent review is recommended fter commencing subcutneous infusion of hydromorphone. The stndrd conversion rtio of 5 mg of morphine to 1 mg of hydromorphone my be n oversimplifiction. A tool is presented tht stndrdizes the ssessment of pin for those ptients unble to complete selfssessment. Acknowledgments The finncil support of the Irish Hospice Foundtion is grtefully cknowledged. Thnks re extended to Mired Mngn for collecting the dt nd to the stff in the hospice for helping with the study. References 1. Cod B, Tnk A, Jcobson RC, Donldson G, Chpmn CR. Hydromorphone nlgesi fter intrvenous bolus dministrtion. Pin 1997;71:41 48. 2. Hys H, Hgen N, Thirlwell M, et l. Comprtive clinicl efficcy nd sfety of immedite relese nd controlled relese hydromorphone for chronic severe cncer pin. Cncer 1994;74:1808 1816. 3. Urquhrt ML, Klpp K, White PF. Ptient controlled nlgesi: comprison of intrvenous versus subcutneous hydromorphone. Anesthesiology 1988;69:428 432. 4. Twycross RG. Pin relief in dvnced cncer. Singpore: Churchill Livingston, 1994:279. 5. Houde RW. Clinicl nlgesic studies of hydromorphone. In: Foley KM, Inturrisi CE, eds. Advnces in Pin Reserch nd Therpy, Vol 8. New York: Rven Press, 1986:129 135. 6. Bbul N, Drke AC, Hgen N, Bker T. Hydromorphone metbolite ccumultion in renl filure. J Pin Symptom Mnge 1995;10:184 186. 7. Bbul N, Drke AC, Hin R. Hydromorphone nd metbolite phrmcokinetics in children. J Pin Symptom Mnge 1995;10:335 337. 8. Moulin DE, Kreeft JH, Murry-Prsons N, Bouquillon AI. Comprison of continuous subcutneous
Vol. 18 No. 1 July 1999 Subcutneous Hydromorphone vs. Morphine 15 nd intrvenous hydromorphone infusions for mngement of cncer pin. Lncet 1991;337:465 468. 9. Twycross RG. Pin relief in dvnced cncer. Singpore: Churchill Livingston, 1994:284. 10. World Helth Orgniztion. Cncer pin relief nd pllitive cre. Report of WHO expert committee. WHO Technicl Report Series 804, Genev: WHO, 1990. 11. Twycross RG. Pin relief in dvnced cncer. Singpore: Churchill Livingston, 1994:255. 12. Fishmn B, Psternk S, Wllenstein SL, Houde RW, Hollnd JC, Foley KM. The memoril pin ssessment crd: vlid instrument for the evlution of cncer pin. Cncer 1987;60:1151 1158. 13. Inghm JM, Portenoy RK. Symptom ssessment. Hemtol Oncol Clin 1996;10:21 39. 14. Ahmedzi S. Recent clinicl trils of pin control: impct on qulity of life. Eur J Cncer 1995; 31A(Suppl 6)S2 S7. 15. Bruer E, Pereir J, Wtnbe S, Belzile M, Kuehn N, Hnson J. Opioid rottion in ptients with cncer pin: retrospective comprison of dose rtios between methdone, hydromorphone nd morphine. Cncer 1996;78:852 857. 16. Lwlor P, Turner K, Hnson J, Bruer E. Dose rtio between morphine nd hydromorphone in ptients with cncer pin: retrospective study. Pin 1997;72:79 85. Appendix 1. Tool for Subjective Assessment of the Ptient Assessment of pin The Memoril Pin Assessment Crd 12 1. Pin Scle LEAST WORST possible possible pin pin 2. Verbl rting scle of 8 pin descriptors no pin, just noticeble, wek, mild, moderte, strong, severe, excruciting 3. Relief Scle NO COMPLETE relief relief of pin of pin 4. Mood Scle WORST BEST mood mood Assessment of dverse effects Indicte the presence or bsence of the following: Yes No Constiption Nuse/vomiting Spontneous jerking of your limbs Buzzing/rumbling in your ers Itch Redness t your syringe driver site Tenderness t your syringe driver site Itch t your syringe driver site Unexpected events (plese indicte below) (Appendix 2 t the following pge.)
16 Miller et l. Vol. 18 No. 1 July 1999 Appendix 2. Tool for Assessment of the Ptient by Proxy Assessment of pin Yes No 1. Is the ptient rigid on turning? 2. Is the ptient moning? - t rest - on turning 3. Is the ptient frowning / grimcing? - t rest - on turning 4. Visul nlogue scle 5. Visul nlogue scle 6. Visul nlogue scle 0 10 s restful s possible s gitted s possible 0 10 s comfortble s possible s uncomfortble s possible no pin 0 10 worst possible pin Assessment of dverse effects Indicte the presence or bsence of the following: Yes No Retching or vomiting Myoclonus Syringe driver site erythem Syringe driver site discomfort Unexpected events (plese indicte below)