Risk Factors for Surgical Shunting in a Large Cohort with Ischemic Priapism

Abstract

Background:

Ischemic priapism is treated with a stepwise algorithm but some patients may benefit from immediate shunt placement.

Aim:

To identify risk factors for surgical shunt placement in a large series of patients with ischemic priapism.

Methods:

We identified all patients presenting to our institution with ischemic priapism from January 2010 to December 2018. Multivariable was performed to assess risk factors for surgical shunting. Receiver operating characteristic (ROC) curve analysis (Youden Index) was used to assess which cutoff time for duration of priapism was most predictive requiring shunting.

Outcomes:

We assess risk factors for surgical shunting and what duration of priapism was most predictive of requiring a shunt.

Results:

We identified a total of 169 ischemic priapism encounters from 143 unique patients, of which 26 (15%) encounters resulted in a surgical shunt. Patients treated with a shunt had longer priapism durations than those without (median 36 vs 10 hours, p<0.001). Independent predictors of a surgical shunt on multivariate logistic regression were duration of priapism in hours (OR 1.05, 95% CI 1.02 to 1.10; p<0.001) and history of prior priapism (OR 3.15, 95% 1.03 to 9.60; p=0.045). ROC curve analysis using priapism duration to predict the need for shunt generated an AUC of 0.83. A duration of 24 hours correlated to a sensitivity of 0.77 and specificity of 0.90.

Clinical Implications:

These results can be used to counsel future patients and assist in the decision-making process for providers

Strengths & Limitations:

This is one of the largest series of priapism in the literature. Most (74%) of the priapism were due to intracavernosal injections so the results may not be generalizable to populations with different priapism etiologies.

Conclusion:

In this study of 169 priapism encounters, we found that priapism duration and history of prior priapism were independent predictors of surgical shunt placement. These results can aid urologists in the counseling and decision-making process of these challenging cases.

Introduction

Priapism is a urologic emergency that can arise from a variety of etiologies. Ischemic priapism is characterized by penile pain and rigidity and can be viewed as a compartment syndrome of the penis as the tissue becomes hypoxic and acidotic.[1] Most cases of priapism are ischemic in nature and need immediate treatment and decompression.[2] If treatment is delayed, the patient may suffer irreversible consequences including smooth muscle necrosis, corporal fibrosis, and permanent erectile dysfunction.[1]

Because priapism is a relatively uncommon disorder, much of the current literature on priapism comes from case reports and smaller case series. The American Urologic Association (AUA) and the European Association of Urology (EAU) both have guidelines on the diagnosis and treatment of priapism. [3][4] While there are small differences between the two, there is a general agreement that treatment for ischemic priapism should consist of a stepwise algorithm starting with corporal irrigation and injection of a sympathomimetic agent. [5] If these maneuvers fail to achieve detumescence, a surgical shunt should then be offered.[5] Several types of surgical shunts have been described.[6] Distal shunts including the Winter, Ebbehoj, Al-Ghorab, and the T-shunt, create a corporoglanular connection while proximal shunts including the Quackels, Sacher, Greyhack, and Barry connect the corpora cavernosa to either the spongiosum of the urethra or a nearby vein. [6]

Currently, shunting is only reserved for patients who do not respond to an attempt at corporal irrigation and injection of sympathomimetics. However, these initial maneuvers, especially when unsuccessful, can delay the time to detumescence, cause significant discomfort, and subject the patient to unnecessary morbidity. Due to the limited data on priapism outcomes, it is difficult to predict which patients would immediately need a shunt on initial presentation. In this study, we use a large retrospective series of patients with ischemic priapism to identify risk factors for surgical shunt placement.

Methods:

The creation of our cohort has been previously described.[7] Briefly, with institutional review board approval, we identified all patients presenting to our institution with priapism from January 1^st, 2010 to July 1^st, 2018. The starting date represents the onset of the Epic® electronic health record system at our institution. Charts were identified with the assistance of the Cedars-Sinai Honest Enterprise Information Service Research Broker (HERB) by an ICD-9 diagnosis code of Priapism (607.3) or an ICD-10 diagnosis code of Priapism (N48.3) including Priapism Unspecified (N48.30), Priapism due to Trauma (N48.31), Priapism due to Disease Classified Elsewhere (N48.32), Drug-Induced Priapism (N48.33), and Other Priapism (N48.39).

Using manual chart review, we extracted information including patient demographics, priapism etiology, priapism duration, and treatments provided. Daily emergency room coverage is split amongst forty attending urologists on staff. There are several emergency room physicians at our institution that will attempt injection of phenylephrine prior to urologic consultation. The type of priapism (ischemic, high flow, stuttering) was also abstracted from the chart review based on the patient’s presentation, symptoms, laboratory studies, and response to treatment. Two cases of stuttering priapism were excluded from further analysis. Priapism etiologies were reviewed by a single physician (H.Z.) and any cases in question were discussed and resolved among two separate physicians (J.A., J.H.). Recreational intracaverosal injections (ICI) were defined as using an intracavernosal agent without a prescription. [7] These were identified when there was documentation of the patient obtaining the injections from a friend or at a party, or if the patient had never seen a physician to obtain the injectables.

We identified all patients that had a surgical shunt placed. The description of the shunt was obtained from the chart review based on the urologists’ notes and operative reports. We abstracted the type and location of the shunt and where the shunt was placed (at bedside or in the operating room).

Univariate (t-test and the Fisher Exact test) and multivariate (logistic regression models) analysis was performed to assess risk factors for surgical shunting. Variables in our model included age, race, history of prior priapism, priapism duration, priapism from all intracavernosal injectable use, priapism from recreational intracavernosal injectable use, corporal irrigation, time of presentation, volume of phenylephrine injected, recreational drug use, and HIV status. Receiver operating characteristic (ROC) curve analysis (Youden Index) was used to assess which cutoff time for duration of priapism was most predictive of requiring shunting.

All analysis was performed with R 3.6.1. Data were considered statistically significant when p<0.05.

Results:

We identified a total of 169 ischemic priapism encounters from 143 unique patients in the 9-year period. Mean age was 45.5 +/− 12.1 years. The etiologies of priapism were recreational ICI (49%), urologist-prescribed ICI (25%), oral PDE5 inhibitors (5%), sickle cell (4%), trazadone (5%), other medication (5%), and unknown (7%). A total of 43 (30%) patients were identified as HIV+. Spontaneous resolution of the priapism occurred in 25 encounters (14.8%). Corporal aspiration and irrigation alone was used in 4 encounters (2.4%) and injection of vasoactive substance alone was used in 19 encounters (11.4%). Both corporal aspiration/irrigation and injection of a sympathomimetic were used in 119 (70.4%) encounters. (Table 1)

Table 1:

Priapism and Patient Demographics

	% (n)
Priapism Encounters	169
Unique Patients	143
Ethnicity	N=143
White	72% (103)
Black	15% (21)
Hispanic	7% (10)
Asian	3% (4)
Other	3% (5)
Priapism Etiology	N=169
Recreational Injectable	49% (82)
Urologist Prescribed Injectable	25% (43)
Oral PDE-5 Inhibitor	5% (9)
Trazadone	5% (9)
Other Medication	5% (9)
Sickle Cell	4% (7)
Unknown	7% (11)
Priapism Treatments	N=169
Urology Consulted	76.9% (130)
Spontaneous Resolution	14.8% (25)
Corporal Aspiration/Irrigation Alone	2.4% (4)
Injection of Sympathomimetic Alone	11.2% (19)
Corporal Aspiration/Irrigation AND Injection of Sympathomimetic	70.4% (119)
Operative Intervention	15.4% (26)
Inpatient Admission	18.3% (31)

There were 26 (15%) encounters that resulted in a surgical shunt placement. Of these, 23 patients (88%) had phenylephrine injection and corporal irrigation prior to the shunt. A total of 8 patients (31%) had a shunt placed at bedside and 18 patients (69%) had a shunt placed in the operating room. The types of shunts are listed in Table 2. None of the patients received a penile prosthesis during the same admission.

Table 2:

Types of Shunts

Bedside	n=8
T-Shunt	5
Ebbehoj Shunt	1
Winter Shunt	1
Unspecified	1
Operating Room	18
T-Shunt	6
Ebbehoj Shunt	6
Al-Ghorab Shunt	2
Quackle Shunt	2
Quackle and Al-Ghorab Shunts	2

There were 9 cases of priapism without a specified duration. Of the remaining 160 cases, patients treated with a shunt had longer priapism durations than those without (median 36 vs 10 hours, p<0.001). Multivariate analysis revealed that duration of priapism (OR 1.05, 95% CI 1.02 to 1.10; p<0.001) and history of prior priapism (OR 3.15, 95% 1.03 to 9.60; p=0.045) were independent predictors requiring surgical shunting. (Table 3).

Table 3:

Multivariate Analysis for Probability of Needing a Surgical Shunt

Variable	Odds Ratio	95% CI	p
Patient Age (years)	1.00	0.95 – 1.10	0.14
Race
Asian	Reference	Reference	Reference
Black or African American	0.29	0.02 – 5.10	0.40
Hispanic	0.33	0.01 – 0.60	0.51
White	0.40	0.03 – 5.00	0.47
Other	0.40	0.01 – 24.00	0.66
Prior Priapism	3.15	1.03 – 9.60	0.045
Priapism Duration (hours)	1.05	1.02 – 1.10	<0.001
Use of All Intracavernosal Injectables (ICI)	0.89	0.21 – 3.70	0.87
Use of Recreational ICI	0.71	0.18 – 2.80	0.62
Corporal Irrigation	4.75	0.79 – 28.40	0.09
Time of Presentation	0.71	0.23 – 2.20	0.56
Use of Illicit Drugs	1.75	0.17 – 17.80	0.64
HIV Status	0.84	0.25 – 2.80	0.77

ROC analysis of priapism duration revealed that this variable was highly sensitive and specific for shunting (Area Under Curve: 0.83). (Figure 1). The optimal cutoff time was 24 hours (sensitivity of 0.77 and specificity of 0.90). Only 5% (6/125) of patients who presented with a priapism duration of less than 24 hours received a surgical shunt whereas 57% (20/35) of patients with priapism duration of 24 hours or longer received a surgical shunt.

Figure 1.

Receiver operating characteristic curve of priapism duration demonstrates a high area under the curve.

Discussion:

In this study, we identify risk factors for surgical shunt placement in a large retrospective cohort of patients with ischemic priapism. We find that the duration of priapism and history of prior priapism are independent risk factors for receiving a surgical shunt. Priapism duration was highly sensitive and specific for shunting with an AUC of 0.83. We identified an optimal cutoff time of 24 hours. These findings can be used to counsel future patients and assist in the decision-making process for providers.

The pathophysiology behind ischemic priapism is becoming better understood.[8] There are several molecular mechanisms that can lead to pathologic cavernous smooth muscle relaxation and congestion, including increased levels of adenosine, cAMP, or cGMP, alpha receptor blockade, and microvascular thrombosis.[9] Occlusion of venous outflow from the corpora cavernosa prevents arterial inflow and results in a severely hypoxic, glucopenic, and acidotic environment.[9]^,[10] Over time, permanent damage can occur to the cavernosa. Histological studies have shown that early stages of priapism result in interstitial edema and thickening.[11] After 24 hours of ischemia, there is evidence of smooth muscle necrosis and fibroblast proliferation, and eventually, the endothelial lining becomes destroyed.[11] This tissue destruction with prolonged ischemia appears to have at distinct consequences on smooth muscle function. Cavernosal smooth muscle already has reduced contractility in this ischemic environment. However, when there is irreversible dysfunction and necrosis, the smooth muscle no longer responds to alpha-adrenergic agonists.[12]^,[13] Thus, maneuvers for detumescence including corporal irrigation and instillation of alpha agonists may have no effect and there may be less of a benefit in attempting these initial procedures after 24 hours.[14] However, some will still respond to corporal irrigation and injection of sympathomimetics even after 24 hours, suggesting individual variability in time of onset of tissue damage.[15] In addition, there may be theoretical differences in how effective alpha agonists can treat various priapism etiologies.

Ultimately, the challenge for urologists in this urgent setting is to identify the patients that already have permanent smooth muscle damage. If these patients will not respond quickly to conservative maneuvers, then they need to proceed right to shunt placement. Our series clinically confirms that a priapism duration of 24 hours marks the start of potentially irreversible smooth muscle dysfunction. For men with priapism who already present with a duration beyond this cutoff, it may be worth proceeding directly to shunt placement (after a short trial of irrigation/sympathomimetics) as to not subject the patient to the potential cardiovascular morbidity of high doses of alpha-agonists or the time delay and discomfort in multiple corporal irrigation and aspiration maneuvers. In our cohort, 23 of the 26 patients who received a shunt had undergone corporal irrigation and phenylephrine injection before the shunt. This may suggest potential overuse of these initial maneuvers in patients at higher risk of shunting.

Our multivariate analysis also found that patients with history of priapism were more likely to receive a surgical shunt (OR 3.15, 95% 1.03 to 9.60; p=0.045). These patients may have already experienced some degree of corporal fibrosis and scarring from a prior episode. Thus, they may have reduced blood flow and may not respond as well to corporal irrigation and alpha agonist therapy. In addition, there may be selection bias as providers could be more inclined to treat these cases more aggressively.

Recently, Masterson et al described a cohort of 334 ischemic priapism encounters, of which 33 (9.9%) received a shunt. [15] Using a machine learning algorithm, they were able to correctly identify those needing a surgical shunt in 87.2% of the testing cohort with an AUC of 0.76. [15] However, there was a low sensitivity (40.1%) with their model. [15] In their cohort, sickle cell was the most common etiology (40.7%), and 51% of men had a prior priapism episode. Interestingly, a smaller proportion of patients with sickle cell disease required shunting compared to other etiologies (2.2% vs 15.2%, p=0.035). Although they did not identify a specific priapism duration in their analysis, 27/44 (61%) of patients in their cohort with a duration longer than 36 hours received a shunt and no patients with a duration shorter than 6 hours required a shunt, highlighting the importance of timely intervention for this population. [15] Ultimately, differences in treatment outcomes between their series and ours may reflect differences in the priapism etiologies and differences in treatment patterns between the two centers.

The other significant consequence of prolonged ischemia is permanent erectile dysfunction which can occur in up to 57% of men with priapism.[16] The degree of erectile dysfunction is probably not just related to the duration of the priapism but also the patient’s baseline function and the treatments used to take down the priapism. Surgical shunts in theory could worsen the erectile dysfunction, but by the time the patients need to receive the shunt, they have already had a prolonged period of ischemia and multiple other treatments. Therefore, it is impossible to ascertain which combination of risk factors led to the dysfunction.[17] Nevertheless, because of the limited data on this, it remains important for the clinician to discuss with the patient the potential for future erectile dysfunction regardless of the duration of priapism. As our retrospective study lacked detailed information on baseline erectile function and other patient reported outcomes and had limited follow up, we hope to address this knowledge gap in a future survey of patients in our series.[7] Further studies can also investigate the cost-benefit and various patient reported metrics of immediate shunting for select cases of priapism.

There are several limitations to this study including its retrospective design and single institution experience. Priapism duration was obtained from chart review and may not be entirely accurate. We also have a unique assortment of priapism etiologies, where 74% were due to intracavernosal injections, many of which were not prescribed to the patients using them, and only 4% were attributed to sickle cell disease. [7,18] It is unclear how the etiology of ischemic priapism can impact the response to treatments. Although it was not the goal of this study, we were unable to identify how much time was delayed by unsuccessful corporal irrigation and phenylephrine injection. In addition, we combined proximal and distal shunts as one outcome for the analysis. Although they both fall under the category of surgical shunting, these two types of shunts can have very different indications and morbidities. Nevertheless, this is one of the largest series of priapism encounters in the literature to our knowledge and encompasses a large number of urologic providers. Further prospective studies and multi-institutional collaborations are needed to continue to improve the treatment algorithm for priapism.

Conclusion:

In this study of 169 priapism encounters, we found that priapism duration and history of prior priapism were independent predictors of surgical shunt placement. These results can aid urologists in the counseling and decision-making process of these challenging cases. The optimal cutoff time on ROC analysis was 24 hours (sensitivity of 0.77 and specificity of 0.90). In order to minimize the morbidity and time delay associated with corporal irrigation and injection of sympathomimetics, urologists may consider proceeding to shunt placement for cases of priapism that present with a duration longer than 24 hours or for patients with a past history of priapism.