A review of outcomes of an intracavernosal injection therapy programme


John Mulhall, Sidney Kimmel Center for Prostate and Urologic Cancers, 353 East 68th Street New York NY 10065, USA. e-mail: mulhalj1@mskcc.org


Study Type – Therapy (outcomes research)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Intracavernosal injection (ICI) therapy is an important treatment option for erectile dysfunction. However, high discontinuation rates have been reported for ICI therapy, and a risk of priapism has long been a concern. There has never been a large sample study performed with multivariate analysis to characterise outcomes of ICI therapy.

The present paper reviews ICI therapy outcomes in a very large population of men at a tertiary care Sexual Medicine Clinic over 5 years. Multivariate analysis was used to further characterise these outcomes. The present study shows that for a large percentage of our sample of patients, ICI therapy is a successful treatment strategy. And, while discontinuation rates are still high, many of those not continuing ICI therapy achieved success with phosphodiesterase inhibitors. Also, the incidence of priapism was less in the present study than previously reported.


  • • To review the outcomes, adverse events and discontinuation rates of intracavernosal injection (ICI) therapy in men with erectile dysfunction (ED) in a sexual medicine practice over a 5-year period at a tertiary referral centre.
  • • Since 1983, ICI has become a staple therapeutic option and high success rates have been reported. However, priapism is a significant concern and discontinuation rates are estimated to be >50%.


  • • Men presenting with ED who were enrolled in our ICI programme between September 2002 and August 2006 were followed at least annually.
  • • Patient demographic information, agents used, erectile function outcomes and adverse events were recorded.
  • • Failure was defined as the inability to have penetrative sex. Discontinuation was defined as patient declaration of such, failure to attend an annual follow-up visit or failure to call for a repeat prescription.
  • • Multivariable analysis was used to define predictors of failure to respond to ICI therapy, as well as predictors of discontinuation within 36 months of starting ICI in those patients responding.


  • • In all, 1412 patients had complete data and constituted the study population. Most patients were using Trimix and 89% of Trimix users were capable of having sexual intercourse.
  • • Response rates were lower in pelvic radiation and diabetic patients.
  • • However, the discontinuation rate was significant; it was lower in men who had not undergone radical prostatectomy (RP). Of note, many RP patients discontinued ICI because of recovery of natural or phosphodiesterase type 5 inhibitor-assisted erections.


  • • ICI therapy is associated with very high success rates even in men with high comorbidity profiles; however, the discontinuation rates, even in men who had not undergone RP, by the end of the third year are significant.
  • • Of note, the recorded priapism rate was extremely low (0.5%).

erectile dysfunction


intercavernosal injection


prostaglandin E1


phosphodiesterase type 5 inhibitors


radical prostatectomy


International Index of Erectile Function


Erectile Dysfunction Inventory of Treatment Satisfaction


odds ratio


In 1982, doctors Ronald Virag [1] and Giles Brindley [2] independently conducted the first successful intracavernosal pharmacotherapy for the treatment of erectile dysfunction (ED), with papaverine and phenoxybenzamine, respectively. Their results were published in 1983 and intracavernosal injection (ICI) therapy is now a recognised safe and effective strategy for the treatment of ED [3]. In contemporary series, prostaglandin E1 (PGE1) is the most commonly used agent and compounded mixtures, including phentolamine and papaverine, are commonly used [4–6]

ICI therapy, 25 years after its introduction, can facilitate the vast majority of users to have sexual intercourse [3]. Even since the discovery of phosphodiesterase type 5 inhibitors (PDE5-i), ICI therapy continues to play a significant role in ED treatment [7,8]. Despite its invasiveness, a previous study has shown that patient satisfaction was greater with ICI therapy than PDE5-i in some patients who had tried both [9,10]. Furthermore, ICI plays an important role in penile rehabilitation programmes in patients after radical prostatectomy (RP) [11].

However, ICI therapy has its complications. Priapism is a significant concern as well as pain, ecchymosis, and haematoma formation [3,12]. Discontinuation rates are not insignificant, with rates of >50% over a 5-year period reported in some series. Patients choose to discontinue ICI therapy for various reasons including: insufficient erections, preference for other treatments, needle phobia, patient or partner disinterest, or loss of partner [13]. Despite publications detailing complications and compliance, predicting which patients will experience complications, fail ICI, or discontinue treatment has not been well defined.

The present analysis was undertaken to review outcomes, adverse events and discontinuation rates for ICI therapy in a sexual medicine practice over a 5-year period at a tertiary referral centre.


Patient population: Men presenting with ED between September 2002 and August 2006 who (i) had failed PDE5-i treatment, (ii) had contraindications for PDE5-i, (iii) could not tolerate PDE5-i, or (iv) preferred ICI to PDE5-i were enrolled in the ICI programme. Patient demographic information, agents used, erectile function outcomes and adverse events were recorded.

Injection training and monitoring: A nurse/nurse practitioner trained the patients in our ICI programme over two in-office sessions. They were followed for the first 4–6 at-home injections by telephone to monitor response and to titrate dose for optimal erection hardness and duration. The routine in-office starting dose was 5 units (0.05 mL) of Trimix (papaverine 30 mg/mL, phentolamine 1 mg/mL, and PGE1 10 µg/mL). Bimix (papaverine 30 mg/mL and phentolamine 1 mg/mL) was used for patients reporting pain with Trimix or who needed intracavernosal phenylephrine during training for a penetration hardness erection lasting >1 h. Patients using ‘supertrimix’ (papaverine 30 mg/mL, phentolamine 2 mg/mL and PGE1 20 µg/mL), papaverine alone (30 mg/mL) or PGE1 alone (2.5–40 µg/mL) were also included in the analysis.

Follow-up: Repeat prescriptions for intracavernosal vasoactive agent were not prescribed unless the patient attended follow-up visits. Follow-up was at 4–6 months after programme commencement and then yearly after this. RP patients in the rehabilitation programme were followed every 4–6 months until 24 months after RP. Detailed instructions about priapism were given to the patient with the following guidelines: a penetration hardness erection lasting 2 h required oral pseudoephedrine (120 mg), at 3 h the patient was to call us and at 4 h to be in the Emergency Department. Failure was defined as the inability to have penetrative sex. Discontinuation was defined as patient declaration of such, failure to attend a follow-up visit, or failure to call for a repeat prescription.

Statistics: Multivariable analysis was used to define predictors of failure to respond to ICI therapy and predictors of discontinuation of ICI ≤36 months in patients who responded to this therapy. Factors considered in both models included, patient age, partner age, ethnicity, vasoactive agent dose, cause of ED, duration of ED and medical comorbidity profile.


Patient population: 1412 patients had complete data and constituted the study population. The mean (sd) patient age was 58 (16) years. Of the study population, 46% had diabetes mellitus, 52% hyperlipidaemia, 20% coronary artery disease, 40% RP, and 9% underwent previous pelvic radiation. In all, 7%, 40%, 31%, and 22% of the study population had one, two, three, or four and more vascular risk factors, respectively.

ICI therapy outcomes: Most (84%) patients used Trimix, 13% used Bimix. Papaverine, PGE1, and Supertrimix were each used by 1% of the patients. In all, 89% of users of Trimix were capable of having sexual intercourse. Response rates were lower in pelvic radiation and diabetic patients (Table 1). Table 2 highlights the discontinuation rates in RP and non-RP patients based on duration in the ICI programme. At 36 months after commencement, overall 24% of patients had ceased ICI for reasons other than PDE5-i being effective. The discontinuation rates overall were higher for RP patients; however, a significant proportion (52%) of RP patients ceased ICI therapy because PDE5-i started being effective.

Table 1.  Multivariable analysis of predictors of failure to respond to ICI therapy
VariableOR (95% CI) P
  1. RT, pelvic radiotherapy; DM, diabetes mellitus.

Trimix dose >50 units30.6 (12.2–65.7)<0.001
RP+RT18.8 (11.4–32.6)<0.001
Commencement >12 months after RP3.1 (1.6–5.5)<0.01
DM2.9 (1.8–4.9)<0.01
RP1.8 (1.5–3.9)<0.05
>5-year history of ED1.5 (1.2–4.5)<0.05
Table 2.  Predictors of discontinuation in responders by 36 months
VariableOR (95% CI) P
  1. PD, Peyronie's disease.

RP12.9 (6.0–31.5)<0.001
Trimix dose >50 units7.9 (4.3–13.7)<0.001
Development of PD5.2 (2.3–9.0)<0.001
Non-caucasian4.9 (2.2–7.6)<0.01
Patient age >70 years3.2 (1.8–5.6)<0.01
Partner age >60 years2.3 (1.7–7.2)<0.05
At least one erection >3 h1.4 (1.1–2.1)<0.05
Partner age ≤45 years0.14 (0.07–0.76)<0.001
Patient age ≤45 years0.39 (0.09–0.82)<0.01

Multivariable analyses: The results of this modelling are presented in Tables 1 and 2. Predictors of failure included: high dose Trimix, prior RP especially when combined with radiation therapy and when ICI therapy was started >12 months after RP, the presence of diabetes, and duration of ED of >5 years. For discontinuation, predictors included: high dose Trimix, the development of Peyronie's disease, non-Caucasian ethnicity, older patient age (>70 years more likely to discontinue) and younger patient age (<45 years less likely to discontinue), older partner age (>60 years more likely to discontinue) and younger partner age (<45 years less likely to discontinue), and the presence of any erection of penetration hardness >3 h.

Adverse events: Seven patients had priapism (defined as an erection of penetration hardness of ≥4 h by patient report) at some point during their treatment (0.5%). Of these patients, the mean age was 55 years, six of the seven men were using Trimix, the mean duration of treatment was 10 months, three had had a previous RP, and most importantly six reported self-adjustment of their dose, despite all patients being warned against such a practice.


To our knowledge, this is the first use of a multivariable analysis to define predictors for treatment failure of ICI therapy and predictors for patient discontinuation of treatment.

Almost 90% of patients responded to ICI therapy with a penetration-hardness erection and were capable of sexual intercourse. Predictors for treatment failure as determined by the present analysis are shown in Table 1. A Trimix dose >50 units (0.5 mL) was a highly significant predictor of failure (odds ratio [OR] 30.6, P < 0.001). Our programme initiates all patients at 5 units of Trimix and titrates upward based on patient response. A very high dose of Trimix is an indication of a patient who has shown insufficient or no response to lower doses of Trimix and is indicative of the presence of venous leak. It is our clinical experience that many patients with ED refractory to high doses of Trimix go on to fail ICI treatment and require a discussion concerning penile implant surgery. This analysis statistically confirms it as a risk factor for failure.

RP alone was a significant predictor of treatment failure (OR 1.8, P < 0.05). Varying degrees of ED are experienced by a significant number of men undergoing RP [14–16]. Patients with both combined RP and pelvic radiotherapy were far more likely to fail treatment (OR 18.8, P < 0.001) than all ICI users, as well as those treated with RP alone. This is not entirely surprising as recent data shows high rates of ED in these dual treatment patients [17]. However, the high relative risk as defined by the present analysis is even greater than what might have been expected. Delay between RP and starting treatment with ICI of >12 months predicted failure (OR 3.1, P < 0.01), and affirms our understanding that early treatment after RP may lead to better outcomes [7,8,18,19]. In our post-RP rehabilitation programme, we encourage the commencement of ICI therapy in PDE5-i non-responders as early as 6 weeks after RP.

Diabetes was a significant predictor of failure (OR 2.9, P < 0.5). Diabetes-associated ED involves a multifactorial pathophysiology affecting both vasculature and the nervous system, resulting in irreversible erectile tissue damage [20]. Treatment is likewise complex and must include strict adherence to ED therapy, as well as treatment of diabetes to prevent further damage [21]. Diabetes prevalence continues to rise in the general population and the ED incidence will likely increase with it, making it imperative that primary care physicians educate patients about this complication [22].

For discontinuation, overall 24% of ICI users stopped using ICI because of factors other than a good response to PDE5-i. Half of RP patients ceased ICI but many as a result of PDE5-i response as nerve regeneration occurred with time after surgery. Use of a high dose of Trimix (OR 7.9, P < 0.001) and previous RP (OR 12.9, P < 0.001) were predictors of discontinuation. Both factors are associated with poor erectile response because of venous leak. The development of Peyronie's disease also predicted discontinuation (OR 5.2, P < 0.001). While there is no direct link between the use of ICI and the development of tunical plaque, there is now a link established between RP and the development of Peyronie's Disease. However, many patients have increased anxiety when they develop penile deformity while on ICI therapy and understandably may opt to cease ICI [12].

A very interesting finding in the present study was that patient and partner age predicted discontinuation of ICI therapy. Patients with older partners (>60 years) or advanced age themselves (>70 years) were at a greater risk of discontinuing ICI treatment (OR 2.3, P < 0.05 and OR 3.2, P < 0.01, respectively). Conversely, patients with younger partners or of younger age themselves (both <45 years) were less likely to discontinue treatment (OR 0.14, P < 0.001 and OR 0.39, P < 0.01, respectively). This finding is interesting and further work is necessary to understand exactly why this is the case; however, this may highlight the importance of psychosocial factors in the sexual health of patients. We encourage all patients to have their partners present for at least one of the two ICI training sessions.

The present discontinuation rate is somewhat lower than that reported by Purvis et al. (40%) [23]. In a questionnaire-based study, an inventory was mailed to 1424 patients who completed the office training and home use phases of a penile self-injection [24]. The overall attrition rate was 31% of the 720 men who completed the questionnaire, with a mean follow-up of 38 months. The main reasons for discontinuation were cost of therapy, patient and partner problems with the concept of penile injection, lack of partner availability and spontaneous improvement in erections. In the present analysis, lack of efficacy of therapy was the primary reason for only one of seven discontinuations. Furthermore, adverse effects of penile injections (priapism, penile nodules, pain) appeared to be only minor contributors to discontinuation.

In another analysis, Hsiao et al. [25] previously identified 37 patients with a good erectogenic response to both sildenafil citrate and penile injection therapy. Patients were assessed using the International Index of Erectile Function (IIEF) as well as the Erectile Dysfunction Inventory of Treatment Satisfaction (EDITS). In that study, there was similar erectile function as measured by the IIEF in response to both therapies, but higher treatment satisfaction as measured by the IIEF and EDITS in response to ICI therapy. This was postulated to be the result of consistency of response and duration of rigidity with ICI therapy when compared with sildenafil. Of note, in the present study, those that discontinued included men who failed to return for evaluation and this group may have included patients who sought ICI therapy elsewhere and who may in fact not have ceased ICI. However, it is probable that the number of men in that cohort of patients is small [26].

Patients with any penetration-hardness erection lasting >3 h were more likely to discontinuation (OR 1.4, P < 0.05). Priapism is a well-recognised complication of ICI therapy and ICI is implicated in 25% of cases of all priapism in the adult population [27,28]. Patients in our programme are specifically educated on the dangers of priapism, how to identify it and to promptly report to the Emergency Room if an erection lasts for >3 h, so that they are in the Emergency Room by 4 h. Despite a tightly structured programme and the clear warning to avoid self-adjustment of the dose, some patients opt to ignore this critically important medical advice and this group represents most of our prolonged erection patient population. We rely on patients declaring their erection duration with ICI and it is plausible that the figure of men having erections lasting 2–4 h is not completely reliable. However, the number of men who presented to an Emergency Room is probably much more reliable given the level of anxiety experienced by such patients during the episode and the near certainty that they or the Emergency Room personnel would have been in contact with us.

The published incidence of priapism with ICI therapy ranges from 0.25% to 7.3% [29–31]. We have a 0.5% incidence (and most admitted dose self-adjustment) and this low number can probably be attributed to careful patient selection and excellent education and counselling about ICI, which is a routine part of our programme [30]. Therefore, we can conclude that with good compliance with prescribed treatment regimens and appropriate education on the danger of priapism, ICI is an even safer therapy than previously thought.

The important predictors outlined in the present study have immediate clinical implications. The predictors of failure help us to better address patient expectations for treatment, resulting in a better setting of goals of therapy. Additionally, that treatment delayed >12 months after RP carries a 3.1 relative risk of treatment failure in the present study, validates previous data suggesting that early postoperative treatment leads to better outcomes [19]. Thus, the high rates of patients in the present study achieving penetration-hardness erections (89% and 72% for Trimix and Bimix, respectively) and the low rate of priapism, affirm that ICI therapy is an effective treatment method and even safer than previously thought when patients are provided with the appropriate amount of education, counselling, and follow-up.

The strengths of the present study include: the very large database of patients, thorough multivariate analysis of the data, and regular annual follow-up of its patients. However, a limitation was that data were verbally self-reported by the patients and future studies may benefit from the use of a validated questionnaire. Furthermore, some patients were lost to follow-up for unknown reasons and it is unclear how many stopped ICI therapy because of treatment success with return of natural erectile function or PDE5-i response and how many discontinued because of treatment failure.

In conclusion, ICI therapy is associated with very high success rates even in men with high comorbidity profiles; however, the discontinuation rate, even in patients that have not undergone RP, by the end of the third year is significant. Notably, the recorded priapism rate was extremely low (0.5%). This is thought to be the result of the structured training programme and close at-home monitoring. We have defined predictors of discontinuation and failure.


None declared.