Tarius SAC Tracker®


Results Wire: US FDA Advisory Committee Supports Sprout Pharmaceuticals’ Addyi (Flibanserin) For Hypoactive Sexual Desire Disorder, But With Added Measures To Mitigate Risk – JUN 4, 2015 (BRUDAC-DSRM)


On Thursday, June 4, 2015, the Bone, Reproductive, and Urologic Drugs Advisory Committee (BRUDAC) and the Drug Safety and Risk Management Advisory Committee (DSRM) jointly supported approval of Addyi (flibanserin 100 mg tablets) by Sprout Pharmaceuticals Inc. (Sprout) for the treatment of hypoactive sexual desire disorder (HSDD) in premenopausal women. In particular, 18 members voted “Yes, but only if certain risk management options beyond labeling are implemented.” Six members voted “No.” There were no votes for the other option, “Yes, with labeling alone to manage the risks.”

Key Outcomes of Advisory Meeting

A majority of the Committee, 18 of 24 members, voted that the overall benefit/risk profile of flibanserin is acceptable to support approval for hypoactive sexual desire disorder (HSDD) in premenopausal women, but only if certain risk management options beyond labeling are implemented. The Committee expressed support for the Risk Evaluation and Mitigation Strategy (REMS) proposed by Sprout. In addition, several of the Committee members suggested that the FDA should require additional elements such as elements to assure safe use (ETASU). They specifically recommended the use of prescriber certification and patient certification (as a sort of informed consent). Furthermore, a few of the Committee members also supported pharmacy certification; however, other members of the Committee thought that this could be too burdensome and unduly limit access for patients.

The Committee discussed the clinical significance of the observed placebo-corrected treatment effects of flibanserin on satisfying sexual events, sexual desire, and related distress. The Committee concluded the benefit of flibanserin is only moderate and is only applicable to a subset of HSDD patients. However, the Committee members also weighed the significant unmet need for an approved HSDD treatment.

The Committee discussed their level of concern with the risks of hypotension and syncope when flibanserin is used alone and when flibanserin is used with alcohol. In particular, they discussed whether Sprout’s alcohol interaction study, which was conducted mostly in men who were moderate alcohol drinkers, adequately assesses risk in premenopausal women. The Committee noted that only 2 women were included in the study, and, therefore, they agreed with the FDA that this study fails to reflect the indicated population. As a result, many of the Committee members expressed support for the FDA to require a postmarket dedicated alcohol study to assess the effects of alcohol in the indicated population of premenopausal women. The Committee also supported the inclusion of strong language in the product labeling to warn about the risks associated with using the drug concomitantly with alcohol. One member advocated for a contraindication, and another member recommended a black boxed warning about this risk.

The Committee was asked to take into account the generalizability of the clinical studies to the population of premenopausal women who would likely use flibanserin, if approved. The FDA was concerned that the exclusion of patients taking certain concomitant medications in the first two pivotal trials made the results less applicable to the real-world. The Committee described their level of concern as moderate. Therefore, they emphasized that prescriber and patient education about the risks associated with the use of certain concomitant medications with flibanserin will be important.

Six of 24 members voted that the overall benefit/risk profile of flibanserin is not acceptable to support approval for hypoactive sexual desire disorder (HSDD) in premenopausal women. These members cited the marginal benefit, and they felt that the benefit was not in balance with the risk.

There were 37 speakers in the Open Public Hearing session of the meeting. Of these, 30 speakers were in support of approval, while 7 spoke against approval.

Overall Voting Results

4. VOTE: Is the overall benefit/risk profile of flibanserin acceptable to support approval for hypoactive sexual desire disorder (HSDD) in premenopausal women?

0-A. Yes, with labeling alone to manage the risks

18-B. Yes, but only if certain risk management options beyond labeling are implemented

6-C. No

Individual Voting Results

Member          Vote    Comment

BRUDAC Members:             

Kathyrn Curtis

            B         I voted B, but a somewhat conflicted and still uncomfortable B. There was minimal effect of unclear clinical significance, and I agree with Dr. Orza’s comment that women suffering from HSDD deserve better than this, but the drug did meet the pre-specified endpoints, and at least for about 10% of women it seemed to be clinically meaningful. There were rare SAEs in the target population but unknown rates of those in whatever the actual use population will be. I agree with most of the previous comments that there needs to be strong REMS put into place, and I was a little shocked to learn that we don't really know how REMS work and how to make them most effective, so just in general, I think we need to do a little more work on that. But specifically, I think there do need to be some very strong postmarketing studies, a dedicated alcohol study in women as was mentioned, and some strong actual postmarketing studies on some of the outcomes, not claims data studies, but some studies where we can actually measure the outcomes, and try and better delineate women for whom it will be effective because those are the women, pretty much the only women where the risk-benefit ratio makes sense.

Vivian Lewis (Chair)   B         I voted B, a difficult B. I have the same concerns as everyone, so it was a difficult B because it’s not a terribly effective drug. It’s a modestly effective drug, and safety concerns are significant. I'll add one thing that hasn't really been brought up yet, and that has to do with the effect of people getting drugs through internet pharmacies, and a REMS kind of a plan that would be initiated. I can see it working both beneficially and not beneficially. I mean on the benefit side, it could provide access to say the rural patient who is some distance from a large pharmacy that could have a certified pharmacist who could participate in such a plan. On the risk side, I believe that some patients obtain pharmaceutical products, even those that are supposed to be restricted to prescription only use, through the internet. So many people have computer access; it could provide the opportunity for educational video or another way to introduce them an informed consent process, especially for people who tend to not read everything carefully that's put before them. That's another option for providing patient education about risks and potential benefits.

Amy Whitaker            B         I voted B as well. I think it’s exciting that we'll have a drug in the armamentarium for the treatment of HSDD. Although, I think we all wish that it was a drug that was a better one, but that overall the very modest benefits outweigh the real, but infrequent risks associated with it, as mitigated by a REMS plan. I've said it several times; I'm uncomfortable with overly restrictive ETASUs in general, because I think access is very important. As Dr. Bell-Perkins said, it’s easier to overcome those in urban areas, but there'll definitely be access issues if we are too strict with ETASU or the REMS. I am comfortable with a medication guide and a communication plan, and with rigorous post marketing studies, which have been well outlined by the previous speakers, as well as the limited marketing at first. I do think that there's going to be a huge buzz around this, and that to be able to roll it out slowly and avoid some of the initial commercials, which are going to come down the pike, would probably be a good way to start.

DSRM Members:                  

Kelly Besco    B         I also voted for B. Like other members of the panel, I do have remaining concerns about this medication. I do believe that a black box warning about the interaction with alcohol should be added to the labeling until we have further studies to better understand the interaction, and the effect on the target population. I believe that a REMS program needs to be robust, and that a patient prescriber agreement form should be added to the REMS program so that there is documentation that the patients have been adequately informed of the potential side effects, interactions, and what I don't think has been previously discussed - What course of action to take should they suffer a fall or injury after a syncopal episode?

I agree with establishing a prescriber certification program to ensure this medication is being used appropriately. As far as the pharmacy burden, being a pharmacist myself, in practice and having to respond to these REMS programs, the act of verifying the prescriber certification is quite minimal, and I do not perceive it to be a burden. I would also be in favor of instituting some sort of three month and six month post-initiation audit process that would be conducted through assessment of the patient, and perhaps that would be done through the creation of a patient registry program, so that we could collect additional information on side effects that could be aggregated for further analysis.

Tobias Gerhard           B         I agree, I also voted B, a very difficult B, definitely. Between B and C, I agree with a lot of the comments that were made before. Particularly, with Dr. Stürmer’s suggestions. I think it is quantifying the risks in real world settings, including in realistic situations with, you know, real world use of alcohol and so on, is absolutely critical to enable women and their caregivers to make informed decisions because currently the risk side isn't quantified sufficiently to really allow informed benefit-risk decisions, but the unmet need seems to be so strong that even for a drug with larger, rather modest benefit, I think, you know, approving the product with strong limitations seems to be the right step at this point. I think its very important that these post marketing requirements can't just be, you know, looking at a database four years after the drug has been on the market. This has to be an active registration of patients that will be burdensome and won't be cheap, won't be easy to do. If that turns out to be infeasible and can't be set up, then I'd rather lean towards C than to see this product kind of unregulated on the market without any assurance that we'll, while it is on the market really learn what the true risks are, and what that means for the benefit risk balance.

Jeanmarie Perrone       B         I voted B, but I applaud the people who voted C. I think we've moved the needle at this committee towards marginal drug acceptance using REMS as a tool that shifts the risk benefit towards accepting these and then modifying afterwards. And I hope we’re not in the position in a couple of years of more drug withdraws on that premise, and using a post marketing surveillance, specifically to identify the patients who are being identified already in these studies as having adverse outcomes, and getting a better profile of who's at risk. I also favor prescriber education and perhaps a patient and prescriber registry, as Dr. Phillips suggested.

Marjorie Shaw Phillips           B         I voted B. I’d like to go over some of Dr. Stürmer’s comments about the postmarketing surveillance that's needed. I think its really important to set realistic expectations for those small subset of patients that are pre-menopausal, have HSDD; some of them will get some meaningful benefit from them for themselves, but its not a magical little pink pill, and there's going to be a whole lot of women with sexual dysfunction for whom there's no evidence that it’s going to benefit them. And there are some potential safety concerns that everyone needs to be aware of. I also echo the need for prescriber registry. That is something that I don't think is overly restrictive. As a pharmacist, I don't think there's a benefit to registering the pharmacist or pharmacy. I do think there's a rule for the pharmacist to confirm that it’s a registered provider, an educated provider that's had that discussion with the patient and the patient is knowledgeable before the drug is dispensed. And that is an important safety consideration. I think this is a case where a patient registry would be very valuable to make sure you've captured postmarketing information. One of the things that I think would be very interesting to learn from a postmarketing basis is what the actual success rate and benefit is in both the less-controlled use, as well as the safety in that less-controlled use, and what the discontinuation rate might be. Some of the information that the FDA shared with testosterone - that the large majority of individuals that took it discontinued in less than six months and it wasn't a lifetime therapy - really provided a lot of evidence that it had marginal benefit, and we could very well find that this holds a lot of promise, but not as much benefit for many of the people that might want to try it. I think that's important information for the public to have, along with the long term safety information to make informed decisions going forward.

Til Stürmer      B         I voted B. I was on the fence here. I have to admit. I think this would be a perfect example for something like staggered licensing, but the closest that the FDA has is to vote B, but require postmarketing studies. So, that is my suggestion. The reason to vote B and not C is that there is a clearly an unmet need, and there is proven potential benefit. I think we all agree on that. The overall magnitude of the benefit is not striking, and there is some tendency for the magnitude to be less pronounced in those most severely affected by the condition. I think that is something that needs to be taken into consideration. There are serious safety concerns. So my yes is conditional on the requirement for post-approval study with timely assessment of actual risk for some of the most severe, adverse outcomes: concussions, accidents, including fatal ones, and pregnancy outcomes ... And a large proportion of women treated ... Was the drug, for example, using a registry, but I mean there needs to be more discussion about this. I'm not an expert on REMS, but the REMS with the ETASU as outlined, using informed consent and provider certification, and if needed, a pharmacist certification sounds reasonable to me.

Temporary Members:                       

G. Caleb Alexander     C         I voted no, and I just want to say it’s possible that this product has an untenable risk-benefit profile, so while I will suggest some additional studies, you know, it is possible that no additional studies are going to suffice, according to some people's thresholds. I would like to see a dedicated alcohol study in women. I'd really like to see a pragmatic clinical trial, frankly, that reflects, that better reflects real world populations, though admittedly this is a bit at odds with the FDA's recommendations a year or two ago to essentially stack the deck in favor of the product by studying it among a more selective population where efficacy is more likely to be demonstrated. I have some concerns that people don't appreciate the difficulty of limiting products use to the approved indications and that people may also be vesting more confidence in the REMS program to do so than has been demonstrated to be the case.

If it was approved tomorrow, I think alcohol should be contraindicated, so I guess I've thought a little further about that since an hour or two ago. A variety of DDI's (drug-drug interactions) should go prominently on the product's label, that is drug interactions. I don't know about limiting DTCA (direct to consumer advertising) during the product's early market debut. I mean, I would be in favor of that, but I don't know if there's regulatory precedent for that, and I do think an ETASU should be implemented with restricted circulation, if it were to be approved tomorrow.

Diane Aronson (Patient Rep) C         I also voted no, and for the reasons that have been stated by those who have voted no, and I won't repeat them. But some of them were comments that I made earlier about my concern about safety issues. I also have a concern about the length of time it may take to get postmarketing information because sometimes that's a real lengthy process, so where it’s just been-short term, I have that concern.

Emilia Bagiella C         I voted C. There were several factors that affected my choice. Mainly was my difficulty to really translate the effects side into something that was meaningful from a clinical and clinical trial point of view. The marginal efficacy over a substantial placebo response was too small, in my opinion, to justify the risks that were observed in this trial. In addition to that, the generalizability of the results was another issue, where there was very little evidence that this small effect size would replicate out in a more general population.

I also thought that there were too limited data on the long-term effect of this drug. This is a drug that is going to be a chronic administration. It’s always going to be given, I heard, to some women for the rest of their lives. There's no data whatsoever of what the effects are in the long-term and how it would be possible to assess the long-term effect, in terms of side effect, effect on the outcomes of pregnancies and so on.

I also thought that any REMS would be inadequate at this point, given that there it not enough information, and it would be very difficult in a chronic population to ask women to stop drinking for the rest of their lives, to stop taking some drugs that they might need for the rest of their lives, to not become pregnant because they want to have sex and they can't do the two things together. I think that it would be difficult to assess what the side effects are going to be in the long-term. A woman who has a drop in her blood pressure falls into the tracks and dies there - that's not going to be picked up by any REMS, and we would never know why that happened.

Elizabeth Bell-Perkins (Acting Consumer Rep)         B         I voted B. It was difficult. I think that what's considered modest or minimal is very meaningful for people who have this disorder. I do believe that postmarketing studies should be done following ... You know for safety, pregnancy ... And I mean the whole fertility ... You know, does it effect fertility? the whole piece of it, not just did the pregnancy result in a live birth, for safety for alcohol and what other drugs the patient may be taking, chronically, REMS definitely. Education, and I hope control over the kind of marketing that's done in all mediums: print, TV, internet, social media. It should be welled out in a very careful way, and I would hope that the FDA would have some say over that. Prescriber certification with the pharmacy certification, I like the idea, but on the consumer level, I don't like the idea. To say that, okay, we're going to approve this, and you know, here's some hoops and things you have to jump through and then maybe putting a hoop there that they can't jump through. I come from a rural area and every little extra thing that has to be done is much more magnified than in an urban area, regardless of economic level. So although I like the idea of prescriber certification, I don't know that I'm comfortable with that as far as making it accessible for patients.

Marianne Brandon      B         I voted B, and I am comfortable actually with the sponsors risk management suggestions. It’s not multi-dimensional. I felt like they addressed the safety issues appropriately and will continue to do so. I just want to be conscious of not implementing unnecessary restrictions and not limiting women’s access to this medication.

Kathryn Flynn            B         So, I also voted B. My background is in patient-reported outcomes measurement, and I did not have concerns about the measures that were used for the outcomes, and I thought the demonstrated benefit was clinically significant. I guess, in terms of the REMS in addition to addressing alcohol use, addressing pregnancy would also be something that I would want to make sure was included.

Walid Gellad   B         I voted B also. Just to say a few things, I think if this were the seventh drug in the class, I think it'd be a very different discussion. It was clear from the material that was submitted and the very brave, public comments that there are many women that suffer, and that there are many women for whom the drug will work, and there are many women for whom the drug will not work. The benefits are modest and I use that ... maybe less than modest, but I think that puts it in good company with other approved drugs, unfortunately. (laughter).

What I understand is those benefits are an average across, and there are going to be some people that benefit, and some will not. So, I have serious, serious, serious safety concerns. It is probably clear from my comments.

I think the syncope is really important, even if it’s rare - because it can be serious, it’s unexpected, and you don't know when it’s going to happen, and its potentially accentuated by other things. I think it is not just something to just ignore, and the study was done in a select population under controlled conditions like any randomized trial. When you expand it to the real world, things are going to be worse. So I do think that a REMS is required. I would suggest that prescriber certification would really be the best thing to ensure that patients who need it get it and it’s only given to those who fit the criteria. So, those would be my comments.

*FDA interjection to ask Committee members to give more input on REMS, for those who voted B

I don't have a huge familiarity with the specifics (about REMS), but I think I am comfortable saying this drug can benefit women who are similar to those in the trial who have the diagnosis. I am very uncomfortable saying that there are samples in the office, and someone comes in, like they often do to me and others, saying that they're having issues with libido. It’s not well investigated, drugs are ignored, etc. So, I think, I guess I don't have an answer for you, other than to say that I think that everything should be done to make sure that drug is only given to those who are as similar as possible to the trial. And so if it requires, and one way you can do that is ensure that only those prescribers who see these patients generally would be the ones prescribing them. Sorry for the terrible answer.

*FDA interjection to clarify the type of REMS options they proposed earlier (communication plan, pharmacy and prescriber certification).

Yeah, I don’t want to take anymore time. I think pharmacy certification would be too burdensome. I think prescriber certification is necessary. I think just general educational material will not be enough.

Marsha Guess C         I voted C. I think, again, there are modest benefits but, as I said, I think there are a lot of assumptions that went into those benefits. There's also a number of risks, and I think simple things like not knowing what over-the-counter foods cause CYP inhibitors and things that we have to be educated about if we're going to prescribe these medications (that) are very important so we adorn that post marketing or pre marketing, understanding those things so that providers who are prescribing it can have those things at hand to be able to counsel their patients appropriately.

* The FDA offered the following clarification, “I know that quite a few have recommended prescriber certification without a pharmacy component because of the potential burden, but I want to make the point that in order for us to be able to achieve prescriber certification, to ensure that occurs, that often involves a pharmacy becoming involved. Otherwise, it still becomes somewhat voluntary, unless it’s tied to dispensing the drug.

I might eventually also add some additional clarification. Pharmacy certification doesn't automatically mean that the pharmacists have to counsel the patient in addition to the prescriber counseling. But, yes, if we have prescriber certification, the pharmacist has to verify that the prescriber who wrote the prescription is indeed certified. So, its kind of a one meets, one requires the second step.

I would do the alcohol (interaction study), and then understanding over the over the counter foods and drugs that might effect that syncope and hypotension. Also, by weight there are about 55% who are overweight, so do normal weight people have more syncope or underweight people, really understanding that so we're prescribing it and making sure we're keeping it in people who are going to be the safest.

Julia Heiman   B         Yes, I voted B and on the efficacy side it is ... These are very modest results. On the other hand, modest results can make a lot of difference when you're at a certain point in a clinical problem. And therefore, I'm less concerned about this. So this may at least get something started.

With regard to recommendations, I think that ... So, I'm trying to balance burden on the entire health care system for patients from docs through pharmacists to what still seems to be needed because its unknown. So, at least for the moment, what I would think would be indeed, provide a certification - not pharmacist's certification. It sounds like informed consent would need to go through the pharmacist. I would therefore back off from that. But, I think the checklist idea that someone brought up would be a good substitute and hopefully not too burdensome.

The other thing is postpartum. I would like see more attention to the alcohol issue and a real study done looking at both low dose, like normal dose that people drink, as well as higher dose, not just these really forced challenges of alcohol, and on women of childbearing age. And, also following up longer-term on cancer and pregnancy, particularly cancer.

Crista Johnson-Agbakwu       B         Thanks. I also with my colleagues voted B. And I feel that HSDD is a very real concern, and the findings from the sponsor, I felt that there was clinical significance in the treatment effect. I do, however, think that there is definitely a need for long-term safety studies, especially around pregnancy and women who are seeking to become pregnant.

I think one way to ensure appropriate patient selection, patient education and counseling, is to require some form of physician certification. Whether that involves targeted training and education that's documented, especially in terms of continuing education on this process involving standardized tools, tech classes was mentioned earlier, and a detailed documentation of the informed consent process that is documented in the patient records, would be something that would be important to include. As well as ongoing postmarketing surveillance.

Lorenzo Leggio           B         I voted B as well. I guess for my comments, I was concerned in particular about the alcohol-drug interaction study, But on the other hand, I echo what my colleagues say, that I took under serious consideration, the fact that it’s not the second drug for this disorder, let’s say, but it’s the first drug. So balancing the benefits, even if not very strong on the medical disorder for approval versus the risk, I didn't think that the concern about the alcohol-drug interaction was so serious to lead me to vote for C. So, that's why I voted B. With that in mind, I still have some concern, for which my recommendation would be to do as part of the REMS, the REMS with the provider certification, just like Dr. Gerhard.

A. Michael Lincoff     B         I voted B for reasons that have all been described, but my main focus is that the REMS approach should focus on making sure marketing correctly describes the proportion of patients that should expect a result and correctly highlights those patients who are not having a result should come off therapy. And focus on a strong contraindication against alcohol, as well as provide certification that emphasizes the correct patient selection. There certainly seems to be a lot of patients outside of this narrow diagnostic criteria who would benefit as well, but I think that should be the topic of a future trial, that would also be able to characterize the risk-benefit ratio in those patients. So, I think that it may well help, but it should be studied and not extrapolated.

Phillip HannoC         Yes, I voted C. I really didn't think that the benefits were better to outweigh the risks, and based on that, that's why I made the vote.

Well, I think the alcohol and women ... I mean by big concern was the endpoints. One endpoint didn't really matter how many times you had satisfying sex, and the other endpoint failed on the first two occasions and then was dropped to pick up an endpoint that succeeded on the first two occasions.

So, the whole thing was kind of squishy to me and I just ... and if it was very robust, I would have said yes, this is worth doing, but given the potential for syncope when people are driving and things, you know, days or weeks after they start the medicine and that it’s a chronic medication, I'm not sure there's a whole lot I would recommend that could be done to convince me that it should be approved.

Michele Orza (Acting Consumer Rep)           C         I voted no, and in some degree of agony. I want to make it clear that I do not, for an instant, doubt that the suffering of these women is real. I never doubt anyone's suffering. Nor that they need and deserve treatment. Everyone who is suffering does. But the question I felt we had to answer was whether this is the treatment they need and the treatment they deserve, and my feeling is that they deserve better. I thought that there was enough noise in the data, and the treatment effect was minimal, marginal at best. And I do have some concern that moving forward, that in considering subsequent medications, that FDA will find that the standard is too low and problematic.

I wanted to echo something that another panelist said about the large placebo effect, which does not represent nothing; it represents all of the other things that were done around the drug. Primarily, I imagine simply recognizing that people are suffering and attending to their needs, and I want to make sure that in however the program moves forward, that those components of the treatment are not lost. And I don't have the ... Actually applaud the kinds of endpoints that were looked at here. If you contrast this, for example, with testosterone, the sole endpoint is the blood level of the testosterone, and there are no endpoints considered that patients actually care about, or that matter to patients, and in this case, I think the endpoints that were examined were very meaningful, and that further work should be done in exploring those kinds of endpoints than have us to assess them.

Robert Silbergleit        B         I want to thank the review division. I think that if I'd been on this panel the previous time, I don't think I would have voted the same way that I voted this time. I think the Review Division did an excellent job of requesting the material necessary to make a benefit-risk analysis. You know, we are always faced with not knowing what safety data we don't have, we’d say. So, I do think the postmarketing data in this case is going to be important. I think that a strategy is going to be important because I think that the most likely risk out of this is going to come from physicians who don't use the drug properly because they're not properly educated. So, a strategy that gets physicians the information they need to use it properly is going to be key. Potentially, especially because I have concerns about the marketing of this drug as well, and the physicians might be in a situation where they have to counter a lot of direct to consumer marketing that could lead to misuse of the drug. So, I think that very careful postmarketing enforcement of marketing strategies, as well as REMS, will be important toward the safe use, but think that we finally have sufficient data.

Kevin Weinfurt           B         I voted B, and I’m in agreement with my other B (voting) colleagues here and echo the prior recommendation for what ought to go into the REMS. And I think it’s also worth it to point out, because there's been some confusion, I think, about the type of benefit that was observed. It’s clear to me that there was very consistent benefit on measures we understand and accept well in the field. For some portion of women, some portion of women didn't have those things. So, I think that's impressive. I think that we also are going to be looking forward to postmarketing studies to understand for whom this therapy is most beneficial, and what other risks are present.

And the other thing I wanted to point out, too, was that there were significant improvements that seem to be conferred by other aspects of attending to the patients in these studies through the study protocol. I think that's a design necessity. It helps us understand the pharmacologic effect, but it’s also encouraging to people who are suffering from this to recognize that this could be a part of a broader strategy to confer larger benefit.

METADATA: Sponsor: Sprout Pharmaceuticals Drug Name: flibanserin Drug Class: 5-HT1A/5-HT2A receptor antagonist Indication: hypoactive sexual desire disorder