Posts Tagged ‘androgen suppression’

Hormone Therapy [9.6]

Writer and Curator: Larry H. Bernstein, MD, FCAP

The structure of this article is as follows:

9.6.1 Hormone Treatment Fights Prostate Cancer

9.6.2 Diabetes and Cardiovascular Disease During Androgen Deprivation Therapy for Prostate Cancer

9.6.3 Breast Cancer and Hormone Therapy

9.6.4 Hormone Therapy and Different Ovarian Cancers

9.6.5 Chemotherapy versus hormonal treatment in platinum- and paclitaxel-refractory ovarian cancer: a randomised trial of the German Arbeitsgemeinschaft Gynaekologische Onkologie (AGO) Study Group Ovarian Cancer


9.6.1 Hormone Treatment Fights Prostate Cancer

By R. Morgan Griffin

Hormone therapy for prostate cancer has come a long way in the past few decades. Not so long ago, the only hormonal treatment for this disease was drastic: an orchiectomy, the surgical removal of the testicles.

Now we have a number of medications — available as pills, injections, and implants — that can give men the benefits of decreasing male hormone levels without irreversible surgery.

“I think hormonal therapy has done wonders for men with prostate cancer,” Stuart Holden, MD, Medical Director of the Prostate Cancer Foundation.

Hormone therapy for prostate cancer does have limitations. Right now, it’s usually used only in men whose cancer has recurred or spread elsewhere in the body.

But even in cases where removing or killing the cancer isn’t possible, hormone therapy can help slow down cancer growth. Though it isn’t a cure, hormone therapy for prostate cancer can help men with prostate cancer feel better and add years to their lives.

On average, hormone therapy can stop the advance of cancer for two to three years. However, it varies from case to case. Some men do well on hormone therapy for much longer.

The idea that hormones have an effect on prostate cancer is not new. The scientist Charles Huggins first established this over 60 years ago in work that led to his winning the Nobel Prize. Huggins found that removing one of the main sources of male hormones from the body — the testicles — could slow the growth of the disease.

“This procedure worked dramatically,” says Holden, who is also director of the Prostate Cancer Center at Cedar Sinai Medical Center in Los Angeles. “Before, these men were confined to bed and wracked with pain. Almost immediately afterwards, they improved.”

Huggins found that some types of prostate cancer cells androgens — to grow. Testosterone is one kind of androgen. About 90% to 95% of all androgens are made in the testicles, while the rest are made in the adrenal glands.

Hormone therapy for prostate cancer works by either preventing the body from making these androgens or by blocking their effects. Either way, the hormone levels drop, and the cancer’s growth slows.

In 85% to 90% of cases of advanced prostate cancer, hormone therapy can shrink the tumor.

However, hormone therapy for prostate cancer doesn’t work forever. The problem is that not all cancer cells need hormones to grow. Over time, these cells that aren’t reliant on hormones will spread. If this happens, hormone therapy won’t help anymore, and your doctor will need to shift to a different treatment approach.

There are two basic kinds of hormone therapy for prostate cancer. One class of drugs stops the body from making certain hormones. The other allows the body to make these hormones, but prevents them from attaching to the cancer cells. Some doctors start treatment with both drugs in an effort to achieve a total androgen block. This approach goes by several names: combined androgen blockade, complete androgen blockade, or total androgen blockade.

Here’s a rundown of the techniques.

  • Luteinizing hormone-releasing hormone agonists (LHRH agonists.)These are chemicals that stop the production of testosterone in the testicles. Essentially, they provide the benefits of an orchiectomy for men with advanced prostate cancer without surgery. This approach is sometimes called “chemical castration.” However, the effects are fully reversible if you stop taking the medication.Most LHRH agonists are injected every one to four months. Some examples are Lupron, Trelstar, Vantas, and Zoladex. A new drug, Viadur, is an implant placed in the arm just once a year.

    Side effects can be significant. They include: loss of sex drivehot flashes, development of breasts (gynecomastia) or painful breasts, loss of muscle, weight gain, fatigue, and decrease in levels of “good”cholesterol.

    Plenaxis is a drug that’s similar to LHRH agonists. However, because it can cause serious allergic reactions, it’s not used that often.

  • Anti-androgens. LHRH agonists and orchiectomies only affect the androgens that are made in the testicles. Thus they have no effect on the 5% to 10% of a man’s “male” hormones that are made in the adrenal glands. Anti-androgens are designed to affect the hormones made in the adrenal glands. They don’t stop the hormones from being made, but they stop them from having an effect on the cancer cells.The advantage of anti-androgens is that they have fewer side effects than LHRH agonists. Many men prefer them because they are less likely to diminish libido. Side effects include tenderness of the breasts, diarrhea, and nausea. These drugs are also taken as pills each day, which may be more convenient than injections. Examples are CasodexEulexin, and Nilandron.

    In some cases, starting treatment with an LHRH agonist can cause a “tumor flare,” a temporary acceleration of the cancer’s growth due to an initial increase in testosterone before the levels drop. This may cause the prostate gland to enlarge, obstructing the bladder and making it difficult to urinate. It’s believed that starting with an anti-androgen drug and then switching to an LHRH agonist can help avoid this problem. In patients with bone metastases, this “flare” can lead to significant complications such as bone pain, fractures, and nerve compression.

    Strangely, if treatment with an anti-androgen doesn’t work, stopping it may actually improve symptoms for a short time. This phenomenon is called “androgen withdrawal,” and experts aren’t sure why it happens.

  • Combined Androgen Blockade. This approach combines anti-androgens with LHRH agonists or an orchiectomy. By using both approaches, you can cut off or block the effects of hormones made by both the adrenal glands and the testicles. However, using both treatments can also increase the side effects. An orchiectomy or an LHRH agonist on its own can cause significant side effects like a loss of libido, impotence, and hot flashes. Adding an anti-androgen can cause diarrhea, and less often, nauseafatigue, and liver problems.
  • Estrogens. Some synthetic versions of female hormones are used for prostate cancer. In fact, they were one of the early treatments used for the disease. However, because of their serious cardiovascular side effects, they’re not used as often anymore. J. Brantley Thrasher, MD, a spokesman for the American Urological Association and chairman of urology at the University of Kansas Medical Center, says they’re usually used only after initial hormone treatments have failed. Examples of estrogens are DES (diethylstilbestrol), Premarin, and Estradiol.
  • Other Drugs. Proscar (finasteride) is another drug that indirectly blocks an androgen that helps prostate cancer cells grow. Depending

on the case, doctors sometimes use other anticancer drugs like Nizoral (ketoconazole) and Cytadren (aminoglutethimide.)

  • Orchiectomy. The surgical removal of the testicles was the earliest form of hormone therapy for prostate cancer. However, the procedure is permanent. As with LHRH agonists, side effects can be significant. They include: Loss of sex drive, hot flashes, development of breasts (gynecomastia) or painful breasts, loss of muscle, weight gain, fatigue, and decrease in levels of “good” cholesterol.

Hormone therapy for prostate cancer can cause osteoporosis, which can lead to broken bones. However, treatment with bisphosphonates — like ArediaFosamax, and Zometa — may help prevent this condition from developing.

Hormone (androgen deprivation) therapy for prostate cancer


Hormone therapy is also called androgen deprivation therapy (ADT) or androgen suppression therapy. The goal is to reduce levels of male hormones, called androgens, in the body, or to stop them from affecting prostate cancer cells.

The main androgens are testosterone and dihydrotestosterone (DHT). Most of the body’s androgens come from the testicles, but the adrenal glands also make a small amount. Androgens stimulate prostate cancer cells to grow. Lowering androgen levels or stopping them from getting into prostate cancer cells often makes prostate cancers shrink or grow more slowly for a time. But hormone therapy alone does not cure prostate cancer.

Hormone therapy may be used:

  • If the cancer has spread too far to be cured by surgery or radiation, or if you can’t have these treatments for some other reason
  • If your cancer remains or comes back after treatment with surgery or radiation therapy
  • Along with radiation therapy as initial treatment if you are at higher risk of the cancer coming back after treatment (based on a high Gleason score, high PSA level, and/or growth of the cancer outside the prostate)
  • Before radiation to try to shrink the cancer to make treatment more effective

Several types of hormone therapy can be used to treat prostate cancer. Some lower the levels of testosterone or other androgens (male hormones). Others block the action of those hormones.

Luteinizing hormone-releasing hormone (LHRH) analogs

These drugs lower the amount of testosterone made by the testicles. Treatment with these drugs is sometimes calledchemical castration or medical castration because they lower androgen levels just as well as orchiectomy.

Even though LHRH analogs (also called LHRH agonists or GnRH agonists) cost more than orchiectomy and require more frequent doctor visits, most men choose this method. These drugs allow the testicles to remain in place, but the testicles will shrink over time, and they may even become too small to feel.

LHRH analogs are injected or placed as small implants under the skin. Depending on the drug used, they are given anywhere from once a month up to once a year. The LHRH analogs available in the United States include leuprolide (Lupron®, Eligard®), goserelin (Zoladex®), triptorelin (Trelstar®), and histrelin (Vantas®).

When LHRH analogs are first given, testosterone levels go up briefly before falling to very low levels. This effect is called flare and results from the complex way in which LHRH analogs work. Men whose cancer has spread to the bones may have bone pain. If the cancer has spread to the spine, even a short-term increase in tumor growth as a result of the flare could compress the spinal cord and cause pain or paralysis. Flare can be avoided by giving drugs called anti-androgens for a few weeks when starting treatment with LHRH analogs. (Anti-androgens are discussed further on.)

Degarelix (Firmagon®)

Degarelix is an LHRH antagonist. LHRH antagonists work like LHRH agonists, but they reduce testosterone levels more quickly and do not cause tumor flare like the LHRH agonists do.

This drug is used to treat advanced prostate cancer. It is given as a monthly injection under the skin. The most common side effects are problems at the injection site (pain, redness, and swelling) and increased levels of liver enzymes on lab tests. Other side effects are discussed in detail below.

Abiraterone (Zytiga®)

Drugs such as LHRH agonists can stop the testicles from making androgens, but other cells in the body, including prostate cancer cells themselves, can still make small amounts, which can fuel cancer growth. Abiraterone blocks an enzyme called CYP17, which helps stop these cells from making androgens.

Abiraterone can be used in men with advanced castrate-resistant prostate cancer (cancer that is still growing despite low testosterone levels from an LHRH agonist, LHRH antagonist, or orchiectomy). Abiraterone has been shown to shrink or slow the growth of some of these tumors and help some of these men live longer.

This drug is taken as pills every day. This drug doesn’t stop the testicles from making testosterone, so men who haven’t had an orchiectomy need to continue treatment with an LHRH agonist or antagonist. Because abiraterone also lowers the level of some other hormones in the body, prednisone (a cortisone-like drug) needs to be taken during treatment as well to avoid certain side effects.

Drugs that stop androgens from working


Androgens have to bind to a protein in the cell called an androgen receptor to work. Anti-androgens are drugs that bind to these receptors so the androgens can’t.

Drugs of this type, such as flutamide (Eulexin®), bicalutamide (Casodex®), and nilutamide (Nilandron®), are pills taken daily.

Anti-androgens are not often used by themselves in the United States. An anti-androgen may be added to treatment if orchiectomy, an LHRH analog, or LHRH antagonist is no longer working by itself. An anti-androgen is also sometimes given for a few weeks when an LHRH analog is first started to prevent a tumor flare.

Anti-androgen treatment can be combined with orchiectomy or an LHRH analog as first-line hormone therapy. This is called combined androgen blockade (CAB).

9.6.2 Diabetes and Cardiovascular Disease During Androgen Deprivation Therapy for Prostate Cancer

Nancy L. KeatingA. James O’Malley and Matthew R. Smith
JCO Sep 20, 2006; 24(27):4448-4456

Purpose Androgen deprivation therapy with a gonadotropin-releasing hormone (GnRH) agonist is associated with increased fat mass and insulin resistance in men with prostate cancer, but the risk of obesity-related disease during treatment has not been well studied. We assessed whether androgen deprivation therapy is associated with an increased incidence of diabetes and cardiovascular disease. Patients and Methods Observational study of a population-based cohort of 73,196 fee-for-service Medicare enrollees age 66 years or older who were diagnosed with locoregional prostate cancer during 1992 to 1999 and observed through 2001. We used Cox proportional hazards models to assess whether treatment with GnRH agonists or orchiectomy was associated with diabetes, coronary heart disease, myocardial infarction, and sudden cardiac death. Results More than one third of men received a GnRH agonist during follow-up. GnRH agonist use was associated with increased risk of incident diabetes (adjusted hazard ratio [HR], 1.44; P < .001), coronary heart disease (adjusted HR, 1.16; P < .001), myocardial infarction (adjusted HR, 1.11; P = .03), and sudden cardiac death (adjusted HR, 1.16; P = .004). Men treated with orchiectomy were more likely to develop diabetes (adjusted HR, 1.34; P < .001) but not coronary heart disease, myocardial infarction, or sudden cardiac death (all P > .20). Conclusion GnRH agonist treatment for men with locoregional prostate cancer may be associated with an increased risk of incident diabetes and cardiovascular disease.

9.6.3 Breast Cancer and Hormone Therapy


There are certain hormones that can attach to breast cancer cells and affect their ability to multiply. The purpose of hormone therapy, also called endocrine therapy, is to add, block, or remove hormones.

With breast cancer, the female hormones estrogen andprogesterone can promote the growth of some breast cancer cells. Therefore in some patients, hormone therapy is given to block the body’s naturally occurring estrogen to slow or stop the cancer‘s growth.

There are two types of hormone therapy for breast cancer.

  • Drugs that inhibit estrogen and progesterone from promotingbreast cancer cell growth.
  • Drugs or surgery to turn off the production of hormones from the ovaries.

Faslodex, a estrogen receptor antagonist, binds to estrogen receptors and blocks their effects on cancer cells. Given as an injection, the drug is for HER2-positive metastatic disease in postmenopausal women who have already tried anti-estrogen therapy. Common side effects of Faslodex include:

  • Injection site pain
  • Nausea and vomiting
  • Loss of appetite
  • Weakness, fatigue
  • Hot flashes
  • Cough
  • Muscle, joint, and bone pain
  • Constipation
  • Shortness of breath

Zoladex and Lupron for Breast Cancer

Zoladex and Lupron are drugs that stop the production of estrogen by the ovaries. They are used in premenopausal women for the treatment of estrogen sensitive breast cancer.

Side effects of Zoladex and Lupron include:

  • Fluid retention
  • Hot flashes
  • Irregular menstrual periods
  • Pain at the injection site


Hormone-sensitive breast cancer cells contain proteins known as hormone receptors that become activated when hormones bind to them. The activated receptors cause changes in the expression of specific genes, which can lead to the stimulation of cell growth.

To determine whether breast cancer cells contain hormone receptors, doctors test samples of tumor tissue that have been removed by surgery. If the tumor cells contain estrogen receptors, the cancer is called estrogen receptor-positive (ER-positive), estrogen-sensitive, or estrogen-responsive. Similarly, if the tumor cells contain progesterone receptors, the cancer is called progesterone receptor-positive (PR- or PgR-positive). Approximately 70 percent of breast cancers are ER-positive. Most ER-positive breast cancers are also PR-positive (1).

Breast cancers that lack estrogen receptors are called estrogen receptor-negative (ER-negative). These tumors are estrogen-insensitive, meaning that they do not use estrogen to grow. Breast tumors that lack progesterone receptors are called progesterone receptor-negative (PR- or PgR-negative).

Hormone therapy (also called hormonal therapy, hormone treatment, or endocrine therapy) slows or stops the growth of hormone-sensitive tumors by blocking the body’s ability to produce hormones or by interfering with hormone action. Tumors that are hormone-insensitive do not respond to hormone therapy.

Hormone therapy for breast cancer is not the same as menopausal hormone therapy or female hormone replacement therapy, in which hormones are given to reduce the symptoms of menopause.

Several strategies have been developed to treat hormone-sensitive breast cancer, including the following:

Blocking ovarian function: Because the ovaries are the main source of estrogen in premenopausal women, estrogen levels in these women can be reduced by eliminating or suppressing ovarian function. Blocking ovarian function is called ovarian ablation.

Ovarian ablation can be done surgically in an operation to remove the ovaries (called oophorectomy) or by treatment with radiation. This type of ovarian ablation is usually permanent.

Alternatively, ovarian function can be suppressed temporarily by treatment with drugs called gonadotropin-releasing hormone (GnRH) agonists, which are also known as luteinizing hormone-releasing hormone (LH-RH) agonists. These medicines interfere with signals from the pituitary gland that stimulate the ovaries to produce estrogen.

Examples of ovarian suppression drugs that have been approved by the U.S. Food and Drug Administration (FDA) are goserelin (Zoladex®) and leuprolide (Lupron®).

Blocking estrogen production: Drugs called aromatase inhibitors can be used to block the activity of an enzyme called aromatase, which the body uses to make estrogen in the ovaries and in other tissues. Aromatase inhibitors are used primarily in postmenopausal women because the ovaries in premenopausal women produce too much aromatase for the inhibitors to block effectively. However, these drugs can be used in premenopausal women if they are given together with a drug that suppresses ovarian function.

Examples of aromatase inhibitors approved by the FDA are anastrozole (Arimidex®) and letrozole (Femara®), both of which temporarily inactivate aromatase, and exemestane (Aromasin®), which permanently inactivates the enzyme.

Blocking estrogen’s effects: Several types of drugs interfere with estrogen’s ability to stimulate the growth of breast cancer cells:

  • Selective estrogen receptor modulators (SERMs) bind to estrogen receptors, preventing estrogen from binding. Examples of SERMs approved by the FDA are tamoxifen (Nolvadex®), raloxifene (Evista®), andtoremifene (Fareston®). Tamoxifen has been used for more than 30 years to treat hormone receptor-positive breast cancer.Because SERMs bind to estrogen receptors, they can potentially not only block estrogen activity (i.e., serve as estrogen antagonists) but also mimic estrogen effects (i.e., serve as estrogen agonists). Most SERMs behave as estrogen antagonists in some tissues and as estrogen agonists in other tissues. For example, tamoxifen blocks the effects of estrogen in breast tissue but acts like estrogen in the uterus and bone.
  • Other antiestrogen drugs, such as fulvestrant (Faslodex®), work in a somewhat different way to block estrogen’s effects. Like SERMs, fulvestrant attaches to the estrogen receptor and functions as an estrogen antagonist. However, unlike SERMs, fulvestrant has no estrogen agonist effects. It is a pure antiestrogen. In addition, when fulvestrant binds to the estrogen receptor, the receptor is targeted for destruction.

There are three main ways that hormone therapy is used to treat hormone-sensitive breast cancer:

Adjuvant therapy for early-stage breast cancer: Research has shown that women treated for early-stage ER-positive breast cancer benefit from receiving at least 5 years of adjuvant hormone therapy (2). Adjuvant therapy is treatment given after the main treatment (surgery, in the case of early-stage breast cancer) to increase the likelihood of a cure.

Adjuvant therapy may include radiation therapy and some combination of chemotherapy, hormone therapy, and targeted therapyTamoxifen has been approved by the FDA for adjuvant hormone treatment of premenopausal and postmenopausal women (and men) with ER-positive early-stage breast cancer, andanastrozole and letrozole have been approved for this use in postmenopausal women.

A third aromatase inhibitorexemestane, is approved for adjuvant treatment of early-stage breast cancer in postmenopausal women who have received tamoxifen previously.

Until recently, most women who received adjuvant hormone therapy to reduce the chance of a breast cancer recurrence took tamoxifen every day for 5 years. However, with the advent of newer hormone therapies, some of which have been compared with tamoxifen in clinical trials, additional approaches to hormone therapy have become common (35). For example, some women may take an aromatase inhibitor every day for 5 years, instead of tamoxifen. Other women may receive additional treatment with an aromatase inhibitor after 5 years of tamoxifen. Finally, some women may switch to an aromatase inhibitor after 2 or 3 years of tamoxifen, for a total of 5 or more years of hormone therapy.

Decisions about the type and duration of adjuvant hormone therapy must be made on an individual basis. This complicated decision-making process is best carried out by talking with an oncologist, a doctor who specializes in cancer treatment.

Treatment of metastatic breast cancer: Several types of hormone therapy are approved to treat hormone-sensitive breast cancer that is metastatic (has spread to other parts of the body).

Studies have shown that tamoxifen is effective in treating women and men with metastatic breast cancer (6).Toremifene is also approved for this use. The antiestrogen fulvestrant can be used in postmenopausal women with metastatic ER-positive breast cancer after treatment with other antiestrogens (7).

The aromatase inhibitors anastrozole and letrozole can be given to postmenopausal women as initial therapy for metastatic hormone-sensitive breast cancer (89). These two drugs, as well as the aromatase inhibitor exemestane, can also be used to treat postmenopausal women with advanced breast cancer whose disease has worsened after treatment with tamoxifen (10).

Neoadjuvant treatment of breast cancer: The use of hormone therapy to treat breast cancer before surgery (neoadjuvant therapy) has been studied in clinical trials (11). The goal of neoadjuvant therapy is to reduce the size of a breast tumor to allow breast-conserving surgery. Data from randomized controlled trials have shown that neoadjuvant hormone therapies—in particular, aromatase inhibitors—can be effective in reducing the size of breast tumors in postmenopausal women. The results in premenopausal women are less clear because only a few small trials involving relatively few premenopausal women have been conducted thus far.

No hormone therapy has yet been approved by the FDA for the neoadjuvant treatment of breast cancer.

9.6.4 Hormone Therapy and Different Ovarian Cancers

Lina Steinrud Mørch, Ellen Løkkegaard, Anne Helms Andreasen, Susanne Krüger Kjær, Øjvind Lidegaard
Am J Epidemiol. 2012; 175(12):1234-1242

Postmenopausal hormone therapy use increases the risk of ovarian cancer. In the present study, the authors examined the risks of different histologic types of ovarian cancer associated with hormone therapy. Using Danish national registers, the authors identified 909,946 women who were followed from 1995–2005. The women were 50–79 years of age and had no prior hormone-sensitive cancers or bilateral oophorectomy. Hormone therapy prescription data were obtained from the National Register of Medicinal Product Statistics. The National Cancer and Pathology Register provided data on ovarian cancers, including information about tumor histology. The authors performed Poisson regression analyses that included hormone exposures and confounders as time-dependent covariates. In an average of 8.0 years of follow up, 2,681 cases of epithelial ovarian cancer were detected. Compared with never users, women taking unopposed oral estrogen therapy had increased risks of both serous tumors (incidence rate ratio (IRR) = 1.7, 95% confidence interval: 1.4, 2.2) and endometrioid tumors (IRR = 1.5, 95% confidence interval: 1.0, 2.4) but decreased risk of mucinous tumors (IRR = 0.3, 95% confidence interval: 0.1, 0.8). Similar increased risks of serous and endometrioid tumors were found with estrogen/progestin therapy, whereas no association was found with mucinous tumors. Consistent with results from recent cohort studies, the authors found that ovarian cancer risk varied according to tumor histology. The types of ovarian tumors should be given attention in future studies.


Ovarian cancer is the most lethal of gynecologic cancers. Unfortunately, little is known about its etiology. In recent meta-analyses, investigators have concluded that women taking postmenopausal hormone therapy (HT) have an increased risk of ovarian cancer compared with never users.[1, 2] Two large prospective studies, the Million Women Study and Danish Sex Hormone Register Study, found an overall increased risk of 30%–40%.[3, 4]

Less is known about the association between hormone use and the risk of different histologic subtypes of epithelial ovarian cancer. Other risk factors for ovarian cancer have been found to differ between mucinous and nonmucinous ovarian tumors, supporting the hypothesis of different etiologies.[5, 6]However, previous studies on HT and different types of ovarian tumors were mainly case-control studies, and the numbers of cases were small, especially for mucinous tumors.[1, 7–10] Most prospective cohort studies either did not examine tumor type[1, 4] or had incomplete information on histology.[11]

Recently, Danforth et al.[12] found that estrogen-only therapy (ET) was more strongly associated with the risk of endometrioid tumors than with the risk of other types of epithelial tumors in the Nurses’ Health Study (NHS). The Million Women Study found that with HT use, the highest risk was for serous tumors, whereas there was a lower risk of mucinous tumors.[3] Knowledge about the associations between HTs and subtypes of ovarian cancer will add to the understanding of how HT acts as a promoter of ovarian cancer carcinogenesis. Moreover, if different types of ovarian tumors are to be viewed as separate diseases, that fact should be considered when creating the study designs for future research. Therefore, the aim of the present study was to explore the risks of HT associated with different histologic types of ovarian cancer.

The study cohort was linked to the National Register of Medicinal Product Statistics using participants’ personal identification numbers as the key identifiers. The National Register of Medicinal Product Statistics includes information on the date of the redeemed prescriptions and the specific Anatomical Therapeutic Chemical code, dose, number of packages, defined daily doses, and route of administration (tablet, patch, gel, etc.) The specific Anatomical Therapeutic Chemical codes included in the present study have been described previously.[13]

The information on initiation of HT use (i.e., redeemed prescriptions) was updated daily for each individual during follow-up. The prescribed defined daily doses were used to determine the length of use. We included 4 months after the expiration of the prescription in all records of hormone exposure to account for any delay in recorded diagnoses in Danish registers, prolonged HT use for those taking less than the defined daily dose prescribed, and minor latency time. Thus, gaps between prescriptions of less than 4 month were filled prospectively; that is, a woman was classified as user of the drug at a given point in time if the dispensed supply from the last redemption had not run out or if it had run out within the last m days (where m is the allowed gap length).[14]

Because HT is likely to act as a promoter of ovarian cancer carcinogenesis with a yet unknown latency time, women currently taking hormones were categorized by the regimen that they took for the longest period during the study period. These variables were time varying; that is, if a woman began a new HT regimen, she would be recategorized if and when the time taking that regimen exceeded the amount of time she took the prior categorization HT regimen. The length of use was calculated as the time spent taking all systemic treatments during the study period. Whether a woman had taken hormones before 50 years of age but within the 11-year study period was accounted for in the hormone status categories, and the amount of time for which she took the hormones was accounted for in the duration of use category. The HT categories were HT use (never, past, current nonvaginal HT use, or other current use (i.e., current use of vaginal ET or a hormone intrauterine device)); hormone formulation (ET, estrogen/progestin therapy (EPT), or other (i.e., tibolone, raloxifene, progestin only, or vaginal estrogen)); hormone regimen (cyclic EPT, continuous EPT, or other); route of administration (oral ET, oral EPT or tibolone, dermal ET, dermal EPT, or other); duration of HT in years (never, current, 0.01–4 years, 4.01–7 years, or >7 years or use of vaginal ET or a hormone intrauterine device); and time since last use among former users (never, current, 0.01–2 years, 2.01–4 years, 4.01–6 years, or >6 years or use of vaginal ET or a hormone intrauterine device).

Ovarian Cancer Cases

Until December 31, 2002, we used the Danish Cancer Register to identify cases of primary invasive ovarian cancers and their histologies, using the International Classification of Diseases for Oncologytopography code 183.0 and morphology codes ending with a 3. At time of the present study, information from January 2003 had not been updated in the Danish Cancer Register. Thus, from 2003 onward, the Pathology Register was used for case findings and information on histology. The invasive epithelial tumors were classified as serous (codes M84413, M84603, M84613, and M90143), endometrioid (codes M83803 and M83813), mucinous (codes M84703, M84803, and M90153), clear-cell (codes M83103 and M83133), adenocarcinoma not otherwise specified (code M81403), or epithelial not otherwise specified (codes M80203, M80703, M81303, M85603, M89333, M89803, and M90003). Nonepithelial invasive tumors and borderline tumors were not included. Eight women for whom we did not have histologic information were excluded. Information on the stages of disease was available from the Danish Cancer Register until December 31, 2002.

From 1995 to 2005, a total of 909,946 perimenopausal and postmenopausal women with no previous cancer or removal of ovaries accumulated 7.3 million person-years of observation, corresponding to an average follow-up period of 8.0 years. The number of incident malignant epithelial ovarian cancers during the study period was 2,681. Of these, 1,336 were serous tumors, 377 were endometrioid tumors, 293 were mucinous tumors, 159 were clear-cell tumors, 115 were nonspecified epithelial tumors, and 401 were adenocarcinomas not otherwise specified. At the end of follow up, 63% of the women remained never users of HT, 22% were previous users, and 9% were current users. Compared with never users, hormone users were more likely to have undergone a hysterectomy (18.0% versus 6.2%) or unilateral salpingo-oophorectomy (5.7% versus 1.9%), to have been sterilized (8.4% versus 5.4%), and to be parous (80.8% versus 75.2%). The characteristics of the study population have been published previously.[4]

Compared with never users, current users of hormones had an increased risk of serous tumors (incidence rate ratio (IRR) = 1.7, 95% confidence interval (CI): 1.5, 1.9) and of endometrioid tumors (IRR = 1.7, 95% CI: 1.3, 2.2). Current use of hormones was not associated with the risk of mucinous or clear-cell tumors (Figure 1). The incidence rate ratios for serous ovarian cancer increased with duration of hormone use (0.01–4 years, IRR = 1.5, 95% CI: 1.3, 1.8; 4.01–7 years, IRR = 1.7, 95% CI: 1.4, 2.1; and >7 years, IRR = 2.1, 95% CI: 1.6, 2.8). The incidence rate ratios for other types of epithelial ovarian cancer were not consistently associated with duration of use (Figure 2).

(Enlarge Image)

Figure 1.

Incidence rate ratios of epithelial ovarian cancers associated with current use of hormone therapy, Danish Sex Hormone Register Study, 1995–2005. Values were adjusted for age, period of use, number of births, hysterectomy, sterilization, unilateral oophorectomy or salpingo-oophorectomy, endometriosis, infertility, and educational level. The reference group was never users of hormone therapy (dashed line). Bars, 95% confidence interval.

(Enlarge Image)

Figure 2.

Incidence rate ratios of epithelial ovarian cancers associated with durations of hormone therapy in years, Danish Sex Hormone Register Study, 1995–2005. Values were adjusted for age, period of use, number of births, hysterectomy, sterilization, unilateral oophorectomy or salpingo-oophorectomy, endometriosis, infertility, and educational level. The reference group was never users of hormone therapy (dashed line). Risk estimates for clear-cell cancer are not shown because there were few cases. Bars, 95% confidence interval.

Time Since Hormone Use

We found increased incidence rate ratios for serous ovarian cancers for a period of up to 2 years after cessation of HT. Thereafter, the risk approached that observed in never users. For endometrioid tumors, the risk was not significantly increased after cessation of HT (Figure 3).

(Enlarge Image)

Figure 3.

Incidence rate ratios of serous and endometrioid ovarian cancers associated with time since last hormone therapy use in years, Danish Sex Hormone Register Study, 1995–2005. Values were adjusted for age, period of use, number of births, hysterectomy, sterilization, unilateral oophorectomy or salpingo-oophorectomy, endometriosis, infertility, and educational level. The reference group was never users of hormone therapy (dashed line). Bars, 95% confidence interval.

Estrogen Therapy

Compared with never users, women on unopposed ET had an increased risk of serous tumors (IRR = 1.7, 95% CI: 1.4, 2.1) and a tendency toward an increased risk of endometrioid tumors (IRR = 1.4, 95% CI: 0.9, 2.1). In contrast, the risk of mucinous tumors was decreased (IRR = 0.3, 95% CI: 0.1, 0.8). No association was found between ET and the risk of clear-cell tumors (IRR = 0.6, 95% CI: 0.2, 1.5) (Figure 4).

(Enlarge Image)

Figure 4.

Incidence rate ratios of epithelial ovarian cancers associated with hormone therapy, Danish Sex Hormone Register Study, 1995–2005. A) Estrogen-only therapy; B) estrogen/progestin therapy. Values were adjusted for age, period of use, number of births, hysterectomy, sterilization, unilateral oophorectomy or salpingo-oophorectomy, endometriosis, infertility, and educational level. The reference group was never users of hormone therapy (dashed line). Bars, 95% confidence interval.

Women on oral ET had a statistically significantly increased risk of endometrioid tumors (IRR = 1.5, 95% CI: 1.0, 2.4), and the risks for serous, mucinous, and clear-cell tumors were similar to the risks found for all ET. Because the risk associations between transdermal ET and ovarian cancers were based on a few cases, the data are not shown. Vaginal estrogen alone was associated with an increased risk of serous tumors (IRR = 1.4, 95% CI: 1.1, 1.9), whereas no associations were found with endometrioid, mucinous, or clear-cell tumors (data not shown).

Combined Therapy

Women on combined EPT had increased incidence rate ratios for serous tumors (IRR = 1.6, 95% CI: 1.4, 1.9) and endometrioid tumors (IRR = 2.0, 95% CI: 1.5, 2.6), whereas no associations were found with mucinous or clear-cell tumors (Figure 4). Similar risk associations were found among women on oral EPT. Because there were few cases, data for transdermal EPT are not shown.

Duration of HT

The incidence rate ratios for serous ovarian cancer increased with increased duration of ET and after 7 years reached an incidence rate ratio of 2.9 (95% CI: 1.9, 4.3). The risks for endometrioid ovarian cancer were similar for all durations of ET (Table 1).

Among women on cyclic EPT, the risk of endometrioid ovarian cancer was increased by 70%–140%, whereas the risk was not increased among women on continuous EPT. The risks for serous ovarian cancer were similar regardless of the duration of cyclic or continuous EPT (Table 1). Results from crude and adjusted analyses were almost identical (data not shown).

Stage of Disease

Overall, the associations between HT and risks of different ovarian tumors did not change after adjustment for the stage of disease (Table 2). Although the analyses were slightly weakened by a lower number of cases, the results roughly showed similar incidence rate ratios across the stages of disease (Table 2).


The present large cohort study suggests that there is a differential influence of HT on different subtypes of ovarian cancer. Hormone users had an excess risk of serous and endometrioid tumors but not of mucinous and clear-cell cancers of the ovaries. Both combined EPT and unopposed ET were associated with increased risks of serous ovarian cancer. Furthermore, cyclic EPT and oral ET were associated with increased risks of endometrioid ovarian cancer. In contrast, no HT was associated with risk of clear-cell ovarian cancer, and women who had used ET had a decreased risk of mucinous ovarian cancer.

Serous Ovarian Cancer

Two large prospective cohort studies, the NHS and the Million Women Study, also found an increased risk of serous ovarian cancer among hormone users.[3, 12] In accordance with our finding, the Million Women Study reported an approximately 50% increased risk with HT.[3] The NHS supports our finding that increasing duration of ET is associated with increasing rate ratios for serous ovarian cancer.[12]

Endometrioid Ovarian Cancer

Although the Million Women Study found no association between any HT and the risk of endometrioid ovarian cancer, we found a 70% increased risk.[3] The NHS found a 50% increased risk of endometrioid tumors after 5 years of ET.[12] In our study, women on oral ET had an up to 2-fold increased risk of endometrioid tumors. Because ET increases the risk of endometrial cancer[15] and endometrioid ovarian tumors are histologically similar to endometrial tissue (16), it seems likely that ET acts through similar biologic mechanisms in the development of endometrioid ovarian cancer, a hypothesis suggested by Danforth et al..[12]

Furthermore, the present study suggests that women on cyclic EPT have an increased risk of endometrioid ovarian cancer, whereas the risk is not increased in women on continuous EPT. Only one study addressed the risk of endometrioid ovarian tumors among women on cyclic versus continuous EPT, and those investigators were not able to demonstrate an increased risk with cyclic or continuous EPT.[7] With regard to the development of endometrial cancer, the increased risk has been found to be confined to women on cyclic EPT.[15] Thus, it is possible that cyclic EPT acts through similar biologic mechanisms in the development of endometrioid ovarian cancer.

Mucinous Ovarian Cancer

Compared with women who were never prescribed HT, women on ET had a 70% decreased risk of mucinous ovarian cancer. The Million Women Study also found a decreased risk of approximately 30% with the use of HT.[3] A few other studies have also suggested that HT is associated with a decreased risk of mucinous ovarian cancer.[12, 17, 18] One group of mucinous tumors is similar to endocervical epithelium and another is similar to colonic epithelium.[16] Both HT in general and ET specifically have been found to decrease the risk of colon cancer.[19, 20] It therefore seems plausible that ET could also decrease the risk of mucinous ovarian cancer. Risch et al.[5] were the first to suggest different etiologies for mucinous and nonmucinous ovarian cancers, and a recent Danish study supported this hypothesis by suggesting significant differences in the risk between mucinous and nonmucinous tumors.[6]


Using the same data as in current study, Mørch et al.[4] found a 40% increase in the overall risk of ovarian cancer in current users of hormones, regardless of the duration and type of HT. However, in the present study, the risk of serous ovarian tumors increased with increasing durations of hormone use. This association was more pronounced among women using ET. After 7 years, the risk of serous ovarian cancer had increased 3-fold among women using ET compared with never users. On the other hand, restricting the analysis to mucinous tumors showed a decreased risk among women using ET. Thus, important information about a differential impact of HT, HT types, and associations with duration of hormone use are not described when different ovarian tumors are examined as a combined outcome.

Moreover, the clarification of the different associations between HT and subtypes of ovarian cancer adds to the understanding of how HT acts as a promoter of ovarian cancer carcinogenesis, as the results are in line with the current knowledge about HT-associated risks of cancers with similar epithelial origins. Because of this, it seems plausible that there is a causal association between HT and ovarian cancer. Other risk factors for ovarian cancer differ based on the type of tumor (mucinous vs. nonmucinous), supporting the hypothesis of different etiologies.[5, 6] The differences should be considered in research study design and suggest that different types of ovarian tumors should be viewed as separate diseases.

Strengths of Study

To our knowledge, our nationwide cohort study is the largest conducted thus far to explore the influence of HT on the risk of histologic subtypes of epithelial ovarian cancer. The validity of our outcome is considered to be high, as data from the Cancer Register validated the diagnoses (21–23). The agreement of histologic ovarian cancer diagnoses between the Pathology Register and the Cancer Register is high, and our estimates did not depend on the source of diagnoses.[24] The information on prescribed HT is transferred electronically from all Danish pharmacies by using bar codes, eliminating recall bias. Our information on both exposures and confounders was updated daily through the national registers, making it possible for us to account for changes in exposures. We excluded women with previous cancer because it might affect both the use of hormones and the subsequent risk of ovarian cancer. Our results were adjusted for age, time period, educational level, number of births, and history of hysterectomy, sterilization, unilateral oophorectomy, salpingo-oophorectomy, endometriosis, or infertility. There was, however, no significant confounding by any of the included variables. We found no evidence of earlier detection (surveillance bias) of ovarian tumors among women on HT. Finally, the stage of disease did not bias the differential association between HT and different tumor types.

Limitations of Study

Data from the National Register of Medicinal Product Statistics is not complete for the time period before January 1995. Thus, information about prescriptions for oral contraceptive use was not available for the women in current study who were 50 years of age or older from 1995−2005. Our incidence rate ratios may be underestimated because of confounding by use of oral contraceptives, as oral contraceptive use decreases the risk of ovarian cancer and often leads to HT.[25, 26] We were not able to restrict our analyses to nonobese women. The ovarian cancer risk associated with HT use is probably clearer in nonobese women (i.e., in women with a body mass index, measured as weight in kilograms divided by height in meters squared, <30).[27] Consequently, our results might be underestimated among nonobese. However, the Million Women Study adjusted data for oral contraceptive use, body mass index, age at menopause, alcohol consumption, smoking, and physical activity, and the adjustments did not result in material changes in their estimates.[3] Also, the NHS reported only minimal changes in the association between HT and the risk of ovarian cancer after adjustment for relevant potential confounders, including duration of oral contraceptive use, occurrence of natural menopause, and age at menarche.[12] The lack of information on family history of cancer might have caused an underestimation of risk in our results, as women with a family history of cancer are probably less likely to use hormones. Information on women who underwent surgical procedures was not available in the registers for the oldest women. Hysterectomy and oophorectomy reduce the risk of ovarian cancer and often lead to HT use, probably causing an underestimation of risk in older women in our results. However, despite our uneven adjustment for confounders, the risks for ovarian tumors were nearly identical across age groups.

9.6.5 Chemotherapy versus hormonal treatment in platinum- and paclitaxel-refractory ovarian cancer: a randomised trial of the German Arbeitsgemeinschaft Gynaekologische Onkologie (AGO) Study Group Ovarian Cancer

  1. du Bois, W. Meier, H. J. Lück, G. Emons, V. Moebus, et al.
    Ann Oncol (2002) 13 (2): 251-257

The majority of patients with ovarian cancer are not cured by first-line treatment.Until now, no study could demonstrate any substantial benefit when exposing ovarian cancer patients to second-line chemotherapy. However, most treatment regimens induce toxicity, thus negatively influencing the quality of rather limited life spans. Here we evaluate whether a second-line chemotherapy can offer any benefit compared with a less toxic hormonal treatment. Patients and methods Patients with ovarian cancer progressing during platinum-paclitaxel containing first-line therapy or experiencing relapse within 6 months were eligible. Patients were stratified for response to primary treatment (progression versus no change/response), and measurable versus non-measurable disease. Treatment consisted of either treosulfan 7 g/m2infused over 30 min or leuprorelin 3.75 mg injected subcutaneously or intramuscularly. Both regimens were repeated every 4 weeks. Results This study began in late 1996, and after 2.5 years accrual an interim analysis was performed when several investigators reported their concern about a suspected lack of efficacy. Following this analysis the recruitment was stopped early and the 78 patients already enrolled were followed up. The majority of patients received treatment until progressive disease was diagnosed or death occurred. Treatment delay was observed rarely and dose reduction was performed only in the treosulfan arm in 5% of 150 courses. Overall, both treatment arms were well tolerated. No objective responses were observed. The median survival time was 36 and 30 weeks in the treosulfan and leuprorelin arms, respectively. Overall survival did not differ between patients with relapse 3–6 months after first-line chemotherapy compared with patients with progressive disease within 3 months.

Conclusions The selected patient population represents a subgroup with extremely poor prognosis. Accordingly, results were not impressive. Both treatment arms showed favourable toxicity data, but failed to show remarkable activity, thus adding only limited evidence to the issue of whether patients with refractory ovarian cancer might benefit from second-line chemotherapy. Even stratified analysis did not identify any subgroup of patients in whom the administration of second-line chemotherapy could demonstrate a clinically relevant survival benefit.

Despite the considerable progress that has been achieved in the treatment of advanced ovarian cancer during the last de-cades, the majority of patients are still not cured by first-line treatment. Therefore, development of effective second-line treatment strategies remains a clinically relevant issue. Today standard first-line regimens in many countries contain both paclitaxel and a platinum analogue (e.g. cisplatin [12] or carboplatin [35]). There are only limited data available reporting results gained from second-line therapy following failure of this new first-line regimen. Until now, no guidelines for the selection of second-line treatment regimens have reached universal acceptance [6]. Furthermore, the definitions of recurrent or relapsed disease according to the status of platinum resistance [7] were solely based on data from patients who did not receive the actual standard first-line regimens containing paclitaxel, and therefore have to be re-evaluated. The treatment-free interval, which offers a chance of gaining a benefit from re-treatment with paclitaxel and/or platinum, remains to be defined. However, patients progressing during or relapsing shortly after platinum-paclitaxel probably have a poor prognosis and can be regarded as refractory to both of the drugs they were exposed to. Until now, no study has demonstrated clearly any substantial benefit for these patients when treating them with second-line chemotherapy. However, most treatment regimens induce toxicity, thus negatively influencing the quality of rather limited life spans in this strictly palliative setting. Therefore, the AGO Study Group set about evaluating whether a second-line chemotherapy could offer any benefit compared with a less toxic hormonal treatment.

The decision to use an alkylating agent for second-line chemotherapy was based on the assumption that these agents, which had been part of first-line treatment of ovarian cancer for decades, could offer some benefit as second-line agent after removal from first-line regimens. Treosulfan (Ovastat®, medac, Germany) was chosen as alkylating agent because it has been registered and used frequently in older first-line regimens in Germany, due to a more favourable non-haematological toxicity profile compared with cyclophosphamide [89]. The published data for treosulfan as second-line treatment after platinum failure had been partially contradictory. Two studies using intravenous treosulfan reported response rates of up to 20% in 25 and 72 patients, respectively [1011]. The latter trial included 43 patients with platinum refractory ovarian cancer and showed a 21% response rate. Orally administered treosulfan resulted in response rates of 3, 14 and 19% in 30, 22 and 16 platinum pre-treated patients, respectively [1214]. The only study reporting results of oral treosulfan in platinum refractory patients observed only one response in 30 patients. Therefore, we decided to use intravenous treosulfan as standard chemotherapy arm in this trial.

Leuprorelin (leuproreline acetate; Enantone®, Takeda, Germany), a gonadotropin-releasing hormone (GnRH) analogue, was selected as hormonal treatment in the experimental arm of this study. It could be administered in a similar time schedule as the chemotherapy regimen (monthly injections) and had shown some activity in previously reported studies in platinum pre-treated ovarian cancer. In these trials, leuprorelin had been used either as single agent [1517] or in combination with megestrole acetate or tamoxifen [1819]. Overall, nine responses have been reported in 46 platinum pre-treated patients [cumulative odds ratio (OR) 19.6%; 95% confidence interval (CI) 9% to 34%]. A retrospective review reported higher efficacy for leuprorelin compared with goserelin, thus providing further support for selecting leuprorelin in favour of other GnRH analogues [17]. However, platinum resistance had been reported inconsistently in all these studies, thus leaving some questions unanswered regarding efficiency in this particular group of patients. Toxicity profiles of leuprorelin had been uniformely reported as being mild, making this option potentially useful in this strictly palliative setting. Tamoxifen, another hormonal treatment with an 11% overall response rate reported in a meta-analysis in recurrent ovarian cancer [20], was not selected for this study, because the study group felt that the different mode of application could hamper comparability.

The median observation period was 22.5 months for all patients. The early termination of recruitment resulted in a statistical power of 80% to detect a 20% survival difference (50% versus 69.9%) after 6 months with two-sided testing and an α error of 0.05.

Treatment and tolerability

The majority of patients received treatment until progressive disease was diagnosed or death occurred. The mean and median treatment periods, respectively, were 18 and 16 weeks in the treosulfan arm, and 13 and 10 weeks in the leuprorelin arm. Treatment delay was observed rarely and median intervals per course were 30.8 and 28.6 days in the treosulfan and leuprorelin arms, respectively. Dose reduction was performed only in the treosulfan arm in eight of 150 courses (5%) because of myelosuppression.

Overall, 150 chemotherapy courses and 122 hormonal treatment courses were evaluable for toxicity. Haematological toxicities higher than grade 2 were observed in only a few patients. Thrombocytopenia grade 3/4 occurred in four and one courses in the treosulfan and leuprorelin arms, respectively. Neutropenia grade 3/4 was only observed in one course in each arm and no infections or neutropenic fever was reported. Anaemia greater than grade 2 was observed after seven courses in the treosulfan arm and after two courses in the leuprorelin arm.

Non-haematological toxicities grade 3 or 4 were reported in only few patients in both arms. Treosulfan induced nausea and emesis in 9% of patients compared with 6% of patients in the leuprorelin arm. Hot flushes were reported by one patient in each arm. Three further patients in the treosulfan arm reported grade 3 pain (two patients) and neurotoxicity (one). The latter was due to remaining toxicity from prior platinum-paclitaxel. Alopecia was common but was due to prior treatment. Re-growth of patients’ hair took longer in the treosulfan arm than in the leuprorelin arm. About one-third of patients still had alopecia after treatment with treosulfan compared with 11% in the leuprorelin arm. Fatigue was reported more frequently in the chemotherapy arm (eight of 36 patients versus one of 37 patients, treosulfan versus leuprorelin;P = 0.014, Fisher’s exact test). Overall, both treatment arms were relatively well tolerated resulting in only one treatment cessation due to toxicity.


No objective responses were observed in either of the treatment arms. Disease stabilisation lasting at least 4 weeks (no change) was reported in nine and four patients in the chemotherapy and hormonal treatment arm, respectively. All but one patient showed progressive disease within a median observation period of 22 months. Median progression-free survival was 17 weeks for treosulfan and 10 weeks for leuprorelin (P = 0.035, Wilcoxon test). The difference between both treatment arms remained significant in favour of treosulfan after adjusting for treatment-free interval before study entry (P = 0.028). However, after 6 months only 23% and 14% of patients in the treosulfan and leuprorelin arms had not progressed; corresponding figures for the 12 month observation period were 9% and 5%, respectively (Figure 1).

View larger version:

Figure 1. Progression-free survival (median 17 and 10 weeks for treosulfan and leuprorelin, respectively; P<0.05, log rank test; Kaplan–Meier curves).

At the time of this analysis, seven patients in the treosulfan arm and eight patients in the leuprorelin arm are still alive with disease [hazard ratio (HR) 0.98; 95% CI 0.58–1.67]. The differences observed between the treatment arms did not reach statistical significance (P = 0.87, Wilcoxon test; Figure2). Again, adjusting for a treatment-free interval before study entry did not alter results. The median survival time was 36 and 30 weeks in the treosulfan and leuprorelin arm, respectively. About one-third of patients in each arm were alive after 12 months.

View larger version:

Figure 2. Survival (median 36 and 30 weeks for treosulfan and leuprorelin, respectively; P = 0.87, log rank test; Kaplan–Meier curves).

The impact of time to treatment failure after first-line chemo-therapy on second-line therapy efficacy was analysed bi-categorially. The cut-off was set at 13 weeks, thus comparing patients with progression under first-line therapy or early relapse within 3 months with patients who relapsed 3–6 months after completion of first-line chemotherapy. Overall, the difference between the groups with respect to progression-free survival did not reach statistical significance. Median progression-free survival was 11 and 12 weeks, respectively, for the two groups (P = 0.46, log rank test; HR 0.83; 95% CI 0.51–1.35). Furthermore, overall survival did not differ significantly between patients with relapse 3–6 months after first-line chemotherapy compared with patients with progressive disease within 3 months (P = 0.34, log rank test; HR 0.77; 95% CI 0.46–1.31). However, median survival was slightly superior in the group with a longer progression-free interval after first-line therapy (42 versus 25 weeks). The latter difference did not reach statistical significance and the Kaplan–Meier curves almost fell on top of each other shortly after 1 year of observation (data not shown).

The presence of bi-dimensionally measurable disease had a negative impact on treatment results. Patients with measurable disease showed a median progression-free survival of 11 weeks compared with 19 weeks in patients with non-measurable disease (P = 0.0006, log rank test). Again, overall survival was superior in the group of patients with non-measurable disease, but this difference did not reach statistical significance (median 47 versus 24 weeks; P = 0.18, log rank test). Only 29% of patients with measurable disease compared with 46% of patients with non-measurable disease were alive after 12 months (HR 1.93; 95% CI 0.73–5.16).

Subsequent treatment

In the treosulfan arm, 15 patients received third-line treatments, of whom three were changed over to leuprorelin. The remaining eight patients received: radiotherapy (one), tamoxifen (one) or chemotherapeutic drugs [topotecan (six), etoposide (one), liposomal doxorubicin (one), carboplatin (one), carboplatin-paclitaxel (one)]. Furthermore, 14 patients received fourth-line treatment, including tamoxifen (two), MPA (one), etoposide (two), topotecan (two), and one patient each idarubicin, gemcitabin or mitoxantrone i.p. Finally,three patients received fifth-line cyclophosphamide (one), etoposide (one) or radiotherapy (one). In the leuproreline arm, almost all patients received third-line therapy. Sixteen patients were crossed over to treosulfan, four received intraperitoneal mitoxantrone, two had liposomal doxorubicin and one patient each received etoposide, topotecan, carboplatin, paclitaxel-mitoxantrone or carboplatin-paclitaxel. Two patients received hormonal third-line treatment (one each received tamoxifen and MPA). Fourth-line treatment was offered to seven patients, including radiotherapy (one), topotecan (two), and one patient each liposomal doxorubicin-etoposide, etoposide or etoposide–5-fluorouracil (5-FU). Fifth-line treatment was offered to three patients, including paclitaxel, gemcitabin and 5-FU–platinum. The considerable use of third-line therapies after progression of disease might have hampered survival analysis, which in fact showed no significant difference between the treatment arms (although progression-free survival differed).

Previous SectionNext Section


This study represents a prospectively randomised trial in a very homogenous population. Only patients who were refractory to the standard first-line treatment of advanced ovarian cancer (i.e. platinum plus paclitaxel) were recruited. This selection represents a patient group with an extremely poor prognosis. At the moment, there is only limited evidence that these patients benefit from second-line chemotherapy at all, and more studies in this subgroup are necessary before any recommendations or guidelines can be established.

A randomised trial of the National Cancer Institute of Canada has shown an advantage for one arm over another when comparing 3-weekly topotecan days 1–5 to weekly topotecan in 78 patients, of whom 60% had received prior paclitaxel,and 60% were platinum refractory [23]. This advantage was limited to overall response (23% versus 8%). Progression-free survival differed only at a non-significant level (8 versus 13 weeks), and overall survival did not differ at all. Our trial showed a statistically significant advantage of one arm (treosulfan) with respect to progression-free survival, but failed to show any difference in overall survival. In addition, no differences with respect to response rates were observed. In fact, we did not observe any objective response. The latter could indicate a lack of activity of both study drugs, treosulfan and leuprorelin. However, even higher response rates as reported in the literature did not translate to longer progression-free and overall survival. A prospectively randomised trial comparing liposomal doxorubicin with topotecan included 254 platinum refractory patients; in addition, about two-thirds had received paclitaxel as part of prior therapy [24]. No significant differences were observed in the refractory subgroup: response rates were 7% and 12%, median progression-free survival was 9 and 14 weeks, and median survival was 33 and 37 weeks, respectively. Our observations of median progression free survival of 11 and 17 weeks and median survival of 30 and 36 weeks fit well with the reported data in this poor prognostic subgroup, although we did not observe any objective responses. Another randomised trial in 81 platinum refractory patients comparing paclitaxel with paclitaxel–epirubicin reported response rates of 17% and 34% translating to 2-year survival of 10% and 18% [25]. The corresponding 2-year survival in our trial was 19% and 22%, thus indicating the limited value of objective response rates as predictors for survival or progression-free survival in this poor prognostic subgroup of patients with truely refractory ovarian cancer.

Nevertheless, achieving an objective response might be beneficial in this palliative setting, especially if bulky tumours induce symptoms such as pain or bowl obstruction. However, objective response rates might not sufficiently reflect this potential benefit. Therefore, different response criteria that better reflect the palliative approach in these patients should be evaluated prospectively (e.g. symptom relief, reduction of pain medication or ability of enteral food intake). The development of better tools for the evaluation of genuine second-line chemotherapies becomes even more necessary when taking into account the fact that ovarian cancer becomes more of a chronic disease: mortality does not change substantially, but median and 5-year survival improves, thus indicating a growing need for efficient second-line and higher treatment. These therapies should allow tumour control and simultaneously should not reduce life quality.

This study reports mild toxicity data for both treatment arms, treosulfan and leuprorelin acetate, but, due to the very poor activity levels observed in both arms, adds only limited evidence to the issue of whether patients with refractory ovarian cancer benefit from second-line chemotherapy at all. Even stratified analysis in patients with progressive ovarian cancer versus patients experiencing relapse 3–6 months after first-line therapy, or patients with measurable versus non-measurable diseases, did not demonstrate any subgroup of patients in whom the administration of treosulfan second-line chemotherapy could demonstrate a clinically relevant benefit. Although a very short progression-free interval and the presence of bi-dimensionally measurable disease seemed to turn a bad prognosis into a worse prognosis, none of the differences between the strata showed a consistent and clinically relevant difference in survival. Only progression-free survival was influenced by these factors to some extent. Our data did not indicate that patients with a progression-free interval of >3 months but

However, results were disappointing in all subgroups. A rather small benefit was traded for a higher rate of fatigue in patients receiving chemotherapy. A gain of 6 weeks median progression-free survival in the superior chemotherapy arm in our study and some advantages with respect to response rates in other trials do not convincingly answer the question of whether second-line chemotherapy offers any benefit for patients with refractory ovarian cancer. Further studies are necessary to help to evaluate whether chemotherapy has a role in this subgroup of patients with a very unfavourable prognosis. A randomised comparison between best supportive care and the most active chemotherapy regimen available could theoretically be an appropriate design for such a trial. However, the German AGO investigators did not even broadly accept a randomisation between a hormonal treatment and a chemotherapy arm, as measured by an extremely slow recruitment rate. Furthermore, this study had to be closed prematurely after an interim analysis indicated only very limited activity in both treatment arms. A trial using best supportive care as one treatment arm would probably not be accepted either, although the above-mentioned efficacy data from chemotherapy studies do not provide more optimistic results.

Treosulfan showed an acceptable toxicity profile and at least some activity compared with leuprorelin acetate, thus allowing continuation of clinical research with this alkylating agent. Our subsequent trial in the refractory population compares treosulfan with topotecan (AGO protocol OVAR-2.3) and recruitment is much better, indicating that investigators more easily accept trials comparing two chemotherapy regimens. Quality of life evaluation was included in this protocol in an attempt to improve understanding of the nature of potential gains from second-line therapy.

Besides treosulfan and topotecan, which are further evaluated by our group, several chemotherapy agents have shown some activity in platinum- and paclitaxel-refractory ovarian cancer, and could serve as comparators in pending further trials: ifosfamide [26], hexamethylmelamine [27], gemcitabin [28] and liposomal doxorubicin [23,29]. The difficult task of recruiting large homogenous patient populations to second-line trials supports the ongoing discussions and activities in cooperative groups and networks, such as the worldwide Gynecologic Cancer InterGroup (GCIG), which is already planning and performing intergroup trials of second-line treatment of ovarian cancer.

Read Full Post »

%d bloggers like this: