Vol. 26 No. 5 November 2003 Journal of Pain and Symptom Management 1055

Clinical Note

Hypogonadism and Sexual Dysfunctionin Male Cancer Survivors ReceivingChronic Opioid TherapyArun Rajagopal, MD, Rena Vassilopoulou-Sellin, MD, J. Lynn Palmer, PhD,Guddi Kaur, RN, BSN, and Eduardo Bruera, MDDepartments of Anesthesiology (A.R.), Endocrine Neoplasia and Hormonal Disorders (R.V.-S.),Palliative Care and Rehabilitation Medicine (J.L.P., G.K., E.B.), and Biostatistics (J.L.P.),University of Texas M. D. Anderson Cancer Center, Houston, Texas, USA

AbstractThe purpose of this study was to determine the prevalence of central hypogonadism and sexualdysfunction in male cancer survivors exposed to chronic high-dose oral opioid therapy. Westudied 20 male patients with cancer-related chronic pain who were disease-free for at least oneyear. All patients consumed at least 200 mg-equivalent of morphine on a daily basis for atleast one year. Participants completed the Sexual Desire Inventory questionnaire and serumlevels of testosterone, follicle-stimulating hormone (FSH), and luteinizing hormone (LH) wereassessed. Serum testosterone levels were reduced in these patients. The median value was 140ng/dL (normal 241–827). There was no compensatory increase in FSH and LH. The medianFSH level was 3.5 mIU/mL (normal 1.4–18.1). The median LH level was 2.1 mIU/mL(normal 1.5–9.3). The mean dyadic sexual desire score was 23.9 � 15.7 (normal value,42.8 � 8.9). The mean solitary sexual desire score was 1.3 � 1.9 (normal value,10.6 � 1.9). Our data suggest that chronic exposure to high-dose oral opioid therapy mayresult in marked central hypogonadism and sexual dysfunction. Given the increasing use oflong-term opioid therapy for chronic pain syndromes, further investigation into these findingsis warranted. J Pain Symptom Manage 2003;26:1055–1061. � 2003 U.S. CancerPain Relief Committee. Published by Elsevier Inc. All rights reserved.

Key WordsHypogonadism, cancer pain, opioid, sexual dysfunction

IntroductionMany cancer survivors have chronic pain

from their cancer or cancer-related treatment

Address reprint requests to: Arun Rajagopal, MD, Sectionof Cancer Pain Management, Department of Anes-thesiology, Box 42, U.T. M. D. Anderson CancerCenter,1515 Holcombe Blvd., Houston, TX 77030,USA.Accepted for publication: March 18, 2003.

� 2003 U.S. Cancer Pain Relief CommitteePublished by Elsevier Inc. All rights reserved.

and require opioid analgesics for extendedperiods of time.1 In the non-cancer chronicpain population, there have been several arti-cles in recent years documenting the develop-ment of marked central hypogonadism andsexual dysfunction in patients receiving in-trathecal opioid therapy.2–4 Although intrathe-cal opioid therapy for chronic pain syndromesis becoming more common, for the majority ofpatients the usual route of administration is theoral route. With recent policy initiatives under-scoring the need for aggressive treatment of

0885-3924/03/$–see front matterdoi:10.1016/S0885-3924(03)00331-2

1056 Vol. 26 No. 5 November 2003Rajagopal et al.

pain,5,6 coupled with increased direct-to-con-sumer advertising,7 more consumers are re-questing opioid therapy for their chronicpain syndromes.

Sexual dysfunction is a common problemin cancer survivors. A number of causes havebeen suggested ranging from the effects of ther-apy, negative self-image from the debilitatingeffects of disease or surgery, and chronic fa-tigue, depression, or anxiety.8–11 Unfortunately,sexual dysfunction is often unrecognized by theprimary care provider and patients receivelittle information on how to deal with the prob-lem.8 It has also been shown that sexualproblems relating to cancer treatment do notresolve with time.8,10

Our clinic population includes a group ofpatients who have been free of cancer for anumber of years but have relied on substantialdoses of opioid therapy for chronic cancer-re-lated pain syndromes. We hypothesized thatthese patients may also develop central hypogo-nadism and sexual dysfunction from disruptionof the hypothalamic-pituitary-gonadal axis in amanner similar to non-cancer chronic painpatients exposed to intrathecal opioid therapy.In patients receiving intrathecal therapy, it ispostulated that chronic administration of exog-enous opioids interferes with the role of en-dogenous opioids on pituitary function.4,12

The purpose of this initial study was to deter-mine the prevalence of central hypogonadismand sexual dysfunction in the cancer survivorgroup exposed to high-dose oral opioid therapy.This question has not been previously studied inthe cancer survivor group. Given the increasingrole of the primary care setting in providingmaintenance medications for cancer survivorswith chronic pain, we report on our series of 20patients who have consumed high doses of opi-oids for an extended period of time.

MethodsPatient Selection

We identified 22 active male patients fromour pain clinic population who matched all in-clusion criteria. Twenty adult male patients witha mean age of 50.1 (range 34–77) agreed toparticipate in this cross-sectional survey. Nonehad previously reported problems with sexualfunction. Their cancer status was “no evidence

of disease” and all had chronic pain related totheir cancer therapy for at least one year. Allpatients had been managed for at least oneyear on chronic opioid therapy at a morphine-equivalent daily dose (MEDD) of at least 200mg. Patients with pre-existing conditions thatmight affect the hypothalamic-pituitary-gonadalaxis at the cranial level (e.g., cranial surgery, cra-nial irradiation, pituitary tumors, testosterone re-placement therapy) were excluded from thestudy. Patients with pre-existing conditions thatmight affect the axis at the gonadal level (pelvicradiation, orchiectomy) were included in thestudy since these conditions should not affectthe development of central hypogonadism. Forall patients, the following demographic detailswere recorded: age, marital status, oncologichistory, and current medications. The study wasreviewed and approved by the InstitutionalReview Board. Informed consent was discussedwith and obtained from all subjects.

Sexual Desire Inventory (SDI) QuestionnaireSexual function was assessed using the Sexual

Desire Inventory Questionnaire.13 Assessingsexual dysfunction has always been difficult be-cause of the confounding nature of assessingconsummatory behavior with sexual desire. Inaddition, assessing sexual desire has been diffi-cult because of the confounding nature of as-sessing dyadic desire (engaging in sexual activitywith another person) with solitary desire (en-gaging in sexual behavior by oneself). Assumingthat consummatory behavior necessarily followssome degree of sexual desire, we decided toassess only sexual desire as a basis for sexualfunction. The SDI questionnaire is a validatedtool that specifically assesses sexual desire inthe absence of consummatory behavior and spe-cifically assesses dyadic vs. solitary desire. Nor-mative data exists for 90 healthy college-ageadults.16

Serum AnalysisThe gonadal axis was evaluated by measuring

single serum concentrations of total testoster-one, follicle-stimulating hormone (FSH), andluteinizing hormone (LH). Eighteen out of the20 samples were collected in a 4-hour windowbetween 1000 and 1400. The remaining twowere collected at 1700 and 1715. Although thefree fraction of testosterone has been regardedas the biologically active molecule, we decided

Vol. 26 No. 5 November 2003 1057Hypogonadism and Sexual Dysfunction in Cancer Survivors

to assess total testosterone because studies haveshown that dissociation of protein-bound testos-terone can occur within capillary beds so that theactive fraction is larger than the free fraction.14

In general, the biologically active fraction oftestosterone is free testosterone plus the frac-tion bound to albumin.14 We assumed a normalalbumin in our patients since they were free ofcancer and not undergoing any cancer-relatedtherapy. In addition, although pulsatile secre-tion of FSH and LH has been recognized, wedecided to assess single values because concen-trations are usually within a limited rangearound a mean value. Accordingly, for patientswith a subnormal testosterone we felt that asingle value would be sufficient to distinguishbetween primary gonadal failure with expectedincreased LH and FSH and central hypogo-nadism (either hypothalamic or pituitary with-out elevated FSH/LH).

Statistical AnalysisDescriptive statistics are used to summarize

the data. For the serum analyses, medians arereported instead of means as measures of cen-tral tendency since most variables did not havesymmetric distributions. Although there is anage-related decline in testosterone levels, ingeneral, testosterone levels are relativelyconstant between the ages of approximately 18

and 75,15 an age range that includes all thepatients in our study. After about age 75, there isa fairly rapid decline in circulating andro-gens.15 As such, our institutional laboratorydoes not report age-related normal values.Thus, we did not use distributional statistics toreport testosterone levels.

In the case of the SDI scores, previously pub-lished scores for a cohort of healthy adults arereported as mean value;16 thus, both medianand mean values are reported for compara-tive purposes.

ResultsResults are summarized in Tables 1 and 2.

Eighteen out of 20 patients (90%) had lowlevels of testosterone. The median value of 140ng/dL is more than 100 ng/dL lower than thelow range of normal values. FSH and LH levelswere not elevated; in fact, median values areclearly on the lower side of normal values.

Dyadic and solitary sexual desire values werealso reduced. Our group demonstrated a meandyadic sexual desire score of 23.9 � 15.7 and amean solitary sexual desire score of 1.3 � 1.9.Median values were lower (19.5 for dyadic and0 for solitary), showing that more patients wereat the lower levels. Twelve patients (60%)

Table 1Individual Patient Results

Age Testosterone FSH LHPatient No. (yr) (ng/dL) (mIU/mL) (mIU/mL) Adjuvant Drugs

1 60 110 2.3 1.2 N/A2 55 142 37.9 11.8 N/A3 40 86 1.9 2.6 Gabapentin 900 mg/day4 39 120 3.9 3.7 N/A5 58 98 2.7 0.7 N/A6 44 205 12.0 4.0 N/A7 35 46 0.7 0.5 Gabapentin 900 mg/day8 58 130 1.1 0.7 N/A9 39 216 10.5 2.7 N/A

10 40 190 15.3 3.9 Amitriptyline 125 mg/day11 56 138 2.7 1.4 Gabapentin 3200 mg/day12 60 379 28.6 6.9 N/A13 71 229 4.2 1.2 Nortriptyline 5 mg/day14 59 135 3.0 1.5 N/A15 51 21 5.9 0.5 N/A16 50 132 5.4 4.6 N/A17 46 381 4.0 4.4 Gabapentin 1800 mg/day18 46 221 2.5 2.9 Paroxetine 20 mg/day19 34 203 2.3 0.6 N/A20 62 152 2.2 0.7 Gabapentin 1200 mg/day

1058 Vol. 26 No. 5 November 2003Rajagopal et al.

Table 2Summary Results

Median Range 25% patients 75% patients Normal

Testosterone (ng/dL) 140 21–381 �110 �205 241–827FSH (mIU/mL) 3.45 0.7–37.9 �2.3 �5.9 1.4–18.1LH (mIU/mL) 2.05 0.5–11.8 �0.7 �3.9 1.5–9.3SDI (Dyadic) 19.5 5–53 �9 �32 42.8 � 8.9a

SDI (Solitary) 0 0–5 0 �2 10.6 � 1.9a

aBased on reported data of 90 healthy adults.16

reported a solitary sexual desire score of 0. Al-though this questionnaire has not been pre-viously used in the cancer survivor population,a cohort of 90 healthy adults had a mean scoreof 42.8 � 8.9 for dyadic sexual desire and10.6 � 1.9 for solitary sexual desire.16

In Table 1, we report on adjuvant drugs thatmay affect sexual function such as antidepres-sants and anti-epileptics. Selective serotoninreuptake inhibitors (SSRI) and tricyclic anti-depressants (TCA) are known to cause sexualdysfunction and recent scattered reports havesuggested that gabapentin may also cause sexualdifficulties.17 Five out of 20 patients were takinggabapentin, 2 out of 20 were taking tricyclicantidepressants, and 1 out of 20 was takingparoxetine.

DiscussionOur results suggest that chronic oral opioid

therapy may result in clinically significant cen-tral hypogonadism and sexual dysfunction. Agreat majority of our patients demonstrated apattern of central hypogonadism with low-normal to markedly decreased levels of testos-terone without the expected compensatoryincrease in FSH and LH. Our data are similarto previously published data on chronic ad-ministration of intrathecal opioids.2–4 In thesereports, patients receiving chronic intrathecalopioids developed marked central hypogo-nadism with decreased testosterone and lowto low-normal FSH and LH. The role of endog-enous opioid peptides in regulating gonad-otrophin secretion has been described andreviewed.12 It has been hypothesized that exoge-nous administration of intrathecal opioids mayplay a similar role.4 Our data demonstrate thatchronic usage of high doses of oral opioids mayhave a similar effect in neuroendocrine regula-tion as intrathecal opioids.

The mechanism by which opioids affect hypo-thalamic-pituitary function is not clearly known.Since endogenous opioids regulate gonadotro-phin secretion, there may be a direct effect ofexogenous opioids on hypothalamic-pituitaryfunction as well. However, studies have alsoshown that administration of morphine raisesprolactin levels and it is known that increasedprolactin levels can lead to hypotestoster-onemia.18–20 Thus, the mechanism of centralhypogonadism in this group may be an indirecteffect of hyperprolactinemia.

The data from the SDI demonstrate markedlydecreased sexual desire. Although a strong asso-ciation between sexual desire and free testoster-one levels has been shown in some studies,21,22

there are also studies demonstrating little corre-lation between levels of testosterone and sexualdesire,23,24 thus underscoring the difficulty in-herent in studying such a subjective entity. Inaddition, the reporting of sexual desire, evenin an anonymous questionnaire, undoubtedlycarries a large subjective bias. There aremany variables that can contribute to sexualdysfunction in this population (effects of cancertreatment, loss of physical well-being, negativebody image, drug interactions, etc.) and isolat-ing chronic opioid therapy as an indepen-dent variable is not possible with our data. Twoof our clinic patients (not included in this seriesbecause they were in the process of voluntarilydiscontinuing chronic opioid therapy) have re-ported a marked increase in sexual desire coin-cident with decreasing opioid therapy.Although our data shows a strong associationbetween chronic opioid use and sexual dysfunc-tion, we are unable to comment on a specificcausal relationship between our group’s de-creased sexual desire and consumption ofopioids.

Although certain drugs are known to causesexual dysfunction (e.g., SSRI and TCA), wecannot comment on any causality between our

Vol. 26 No. 5 November 2003 1059Hypogonadism and Sexual Dysfunction in Cancer Survivors

patients’ consumption of these drugs and devel-opment of sexual side effects. Only a smallnumber of our patients were consuming thesedrugs and we were unable to note an associationbetween these drugs and their profoundsexual dysfunction.

LimitationsOur study has limitations. This is an initial

study in this group of patients. Although ourdata have validated our hypothesis that centralhypogonadism and sexual dysfunction are a sig-nificant problem in this group of patients, thisinitial study did not consider a control groupand thus is unable to demonstrate a causal rela-tionship. Since the existence of this problem incancer survivors was unknown prior to this studyand was hypothesized based on data from pa-tients receiving intraspinal opioids, our primarypurpose was only in determining the prevalenceof hypogonadism and not demonstrating cau-sality. In order to definitively demonstratecausality, we would have to withhold opioidtherapy in our patients and assess resumptionof normal gonadal function over a period oftime. Clearly, this would be unjustifiable froman ethical standpoint. In an ongoing study, ourgroup is assessing the prevalence of central hy-pogonadism, sexual dysfunction, depressionand fatigue in a cohort of cancer survivors notconsuming opioids on a chronic basis. Ouractive group is also being assessed for depres-sion and fatigue since these factors may alsoaffect sexual function. Although such a studywill not demonstrate causality per se, identi-fying normal gonadal and sexual function anddecreased depression and fatigue in cancersurvivors who are not taking chronic opioidtherapy will strengthen our argument.

Our choice of the SDI for evaluation of sexualdesire was based on the belief that a centralmechanism for hypogonadism would lead toreduced sexual desire. Although better compar-ison studies have been reported, especially withthe pharmaceutical studies assessing sildenafilfor patients with erectile dysfunction,25,26 thesereports largely tend to emphasize sexual func-tion rather than desire. We believe that studiesmeasuring sexual function, which assume thepresence of a partner, may confound the assess-ment of whether opioids specifically affect inter-est in sexual function. The SDI questionnairehas been validated only in a population of

young, college-age, healthy adults. Thus, a com-parison between this group and our patientpopulation is valid only to demonstrate the pro-found sexual dysfunction that exists in our pop-ulation. Our ongoing study, which assessescancer survivors not consuming opioids on achronic basis, should offer a better comparisongroup.

Although we excluded from our study pa-tients whose hypothalamic-pituitary axis mightbe affected by prior conditions relating to theironcologic history (e.g., cranial radiation, pitu-itary tumors, cranial surgery), central hypogo-nadism has many causes. In this initial study,we did not attempt to identify causes unrelatedto their oncologic history. Thus, it is possiblethat some of our patients had pre-existing hypo-gonadism unrelated to their chronic opioidhistory.

Finally, in this selected group of high-doseopioid patients, our data cannot establish adose-response effect between daily opioid doseand development of secondary hypogonadism.It is possible that this apparent hypogonadicstate is induced only in patients receiving high-dose therapy. A recent report demonstrated alower prevalence of hypogonadism in patientsreceiving lower doses of opioids.27 Future re-search should address this question.

ConclusionOur data demonstrate that the prevalence of

sexual dysfunction and central hypogonadismin this group of patients is very high and corre-lates with published data on the chronic painpatient receiving intrathecal opioids. Sincechronically low levels of testosterone can lead toproblems such as chronic anemia, decreasedbone mass, muscle atrophy and even cognitiveimpairment,28,29 our data suggest that testoster-one replacement may be indicated in some pa-tients receiving chronic opioid therapy.

Future research in this area should includean assessment of concurrent fatigue and depres-sion and the role of chronic opioid use in thefemale population since this group is addition-ally at higher risk of development of osteoporo-sis. Although overall concentration is lower inwomen, testosterone has an important phy-siological role in women and chronic hypogo-nadism from opioid use may lead to the same

1060 Vol. 26 No. 5 November 2003Rajagopal et al.

problems mentioned above.30–33 With the in-creased emphasis being placed on pain treat-ment and the expansion of chronic opioidtherapy into primary care settings, furtherstudies into these and other related areas arewarranted.

References1. Foley K. The treatment of cancer pain. N Engl

J Med 1985;313:84–95.

2. Paice JA, Penn RD, Ryan WG. Altered sexualfunction and decreased testosterone in patients re-ceiving intraspinal opioids. J Pain Symptom Manage1994;9:126–131.

3. Abs R, Verhelst J, Maeyaert J, et al. Endocrineconsequences of long-term intrathecal administra-tion of opioids. J Clin Endocrinol Metab 2000;85:2215–2222.

4. Finch PM, Roberts LJ, Price L, et al. Hypogo-nadism in patients treated with intrathecal mor-phine. Clin J Pain 2000;16:251–254.

5. Cleeland CS, Cleeland LB, Dar R, et al. Factorsinfluencing physician management of cancer pain.Cancer 1986;58:796.

6. Von Roenn JH, Cleeland CS, Gonin R, et al.Physician attitudes and practice in cancer pain man-agement. A survey from the Eastern CooperativeOncology Group. Ann Intern Med 1993;119:121–126.

7. Kopp SW, Bang HK. Benefit and risk informationin prescription drug advertising: review of empiricalstudies and marketing implications. Health Market-ing Quarterly 2000;17:39–56.

8. McKee AL, Schoever LR. Sexuality rehabilita-tion. Cancer 2001;92:1008–1012.

9. Schover LR, Montague DK, Lakin MM. Sexualproblems. In: DeVita VT, Hellman S, Rosenberg SA,eds. Cancer: principles and practice of oncology,5th ed. Philadelphia: Lippincott-Raven, 1997:2857–2872.

10. Litwin MS, Hays RD, Fink A, et al. Quality-of-lifeoutcomes in men treated for early prostate cancer.JAMA 1995;273:129–135.

11. Arai Y, Kawakita M, Okada Y, et al. Sexuality andfertility in long-term survivors of testicular cancer. JClin Oncol 1997;15:1444–1448.

12. Genazzani AR, Genazzani AD, Volpogni C, et al.Opioid control of gonadotrophin secretion inhumans. Hum Reprod 1993(Suppl 2);8:S151–153.

13. Spector IP, Carey MP, Steinberg L. The SexualDesire Inventory: Development, factor structure, andevidence of reliability. J Sex Marital Ther 1996;22:175–190.

14. Pardridge WM. Serum bioavailability of sex ste-roid hormones. Clin Endocrinol Metab 1986;15:259–278.

15. Griffin JE, Wilson JD. Disorders of the testis andthe male reproductive tract. In: Wilson JD, FosterDW, Kronenberg HM, et al., eds. Williams Textbookof Endocrinology, 9th ed. Philadelphia: WB Saun-ders, 1998:831.

16. King BE, Allgeier ER. The Sexual Desire Inven-tory as a measure of sexual motivation in collegestudents. Psychol Rep 2000;86:347–350.

17. Husain AM, Carwile ST, Miller PP, et al. Im-proved sexual function in three men taking lamotrig-ine for epilepsy. Southern Medical Journal 2000;93:335–336.

18. Delitala G, Grossman A, Besser GM. The partici-pation of hypothalamic dopamine in morphine-in-duced prolactin release in man. Clin Endocrinol(Oxf) 1983;19(4):437–444.

19. Zis AP, Haskett RF, Albala AA, et al. Morphineinhibits cortisol and stimulates prolactin secretion inman. Psychoneuroendocrinol 1984;9:423–427.

20. Su CF, Liu MY, Lin MT. Intraventricular mor-phine produces pain relief, hypothermia, hyperglyce-mia and increased prolactin and growth hormonelevels in patients with cancer pain. J Neurol 1987;235:105–108.

21. Lundberg PO, Hulter B. Sexual dysfunction inpatients with hypothalamo-pituitary disorders. ExpClin Endocrinol 1991;98:81–88.

22. Ahn HS, Park CM, Lee SW. The clinical rele-vance of sex hormone levels and sexual activity inthe aging male. BJU International 2002;89:526–530.

23. Schiavi RC, White D, Mandeli J, et al. Effect oftestosterone administration on sexual behavior andmood in men with erectile dysfunction. Arch SexBehav 1997;26:231–241.

24. Guay AT, Bansal S, Heatley GJ. Effects of raisingendogenous testosterone levels in impotent menwith secondary hypogonadism: double blind pla-cebo-controlled trial with clomiphene citrate. J ClinEndocrinol Metab 1995;80:3546–3552.

25. Rosen RC, Riley A, Wagner G, et al. The interna-tional index of erectile function (IIEF): a multidi-mensional scale for assessment of erectiledysfunction. Urology 1997;49:822–830.

26. Rosen RC, Cappelleri JC, Gendrano N, 3rd. TheInternational Index of Erectile Function (IIEF): astate-of-the-science review. Int J Impot Res 2002;14:226–244.

27. Daniell HW. Hypogonadism in men consumingsustained-action oral opioids. J Pain 2002;3:377–384.

28. Zitzmann M, Nieschlag E. Hormone substitutionin male hypogonadism. Mol Cell Endocrinol 2000;161:73–88.

29. Morales A, Heaton JP. Hormonal erectile dys-function. Evaluation and management. Urol ClinNorth America 2001;28:279–288.

Vol. 26 No. 5 November 2003 1061Hypogonadism and Sexual Dysfunction in Cancer Survivors

30. Carlson ES, Li S. Androgen therapy for meno-pausal women. Clinical Excellence for Nurse Prac-titioners 1998;2:324–328.

31. Mazer NA. Testosterone deficiency in women:etiologies, diagnosis, and emerging treatments. IntJ Fertility & Women’s Medicine 2002;47:77–86.

32. Modelska K, Cummings S. Female sexual dys-function in postmenopausal women: systematicreview of placebo-controlled trials. Am J Obstet Gyn2003;188:286–293.33. Guay AT. Decreased testosterone in regularlymenstruating women with decreased libido: a clinicalobservation. J Sex & Marital Ther 2001;27:513–519.