Opioid Pain Medication Use After Dermatologic Surgery: A Prospective Observational Study of 212 Dermatologic Surgery Patients

Objective To better understand postoperative opioid use after dermatologic surgery.

Design Prospective observational study.

Setting Academic dermatology department.

Patients The study included 212 adults (1) who were undergoing a single skin excision (including Mohs micrographic surgery), (2) who consented to participate, and (3) who were able to be reached by telephone on postoperative day 3 or 4. Patients who did not meet these criteria and those referred to another physician for further surgical treatment or repair were excluded.

Main Outcome Measures The study examined (1) the incidence of opioid prescription after dermatologic surgery, (2) the percentage of prescribed opioid pain medications used in the postoperative period, and (3) patient and surgical characteristics associated with opioid pain medication prescription and use.

Results Opioids were prescribed to 72 of the 212 patients (34%). Twenty-five of the 72 patients (35%) who were prescribed opioids did not use them. Forty-nine of 57 patients (86%) who filled an opioid prescription had leftover pills, and 26 of the 49 patients (53%) planned to keep them. Only maximum pain score was significantly associated with opioid use.

Conclusions Opioids were overprescribed after dermatologic surgery. Patients who had leftover opioids did not dispose of them properly, which could lead to potential misuse and abuse.

Opioids are potent pain-reducing agents and are used frequently after surgical procedures, including dermatologic surgery. Per capita opioid sales in the United States increased 402% between 1997 and 2007.¹ There has been a corresponding increase in opioid misuse and abuse, with a 3-fold increase in opioid poisoning deaths between 1999 and 2006.² In Utah, the state in which we practice, prescription drug overdose resulted in more deaths than motor vehicle crashes in 2009, and most overdose deaths involved opioids.³ Furthermore, in a recent Utah telephone questionnaire, 21% of adults reported receiving an opioid prescription in the past year, and 72% of respondents who received an opioid prescription reported having leftover opioid medication, suggesting that opioid pain medications may frequently be overprescribed.⁴

Dermatologic operations are common procedures, yet data on the postoperative use of opioid pain medication after dermatologic surgery are limited. To our knowledge, only 1 study has reported on the frequency of opioid use after dermatologic surgery; it reported that 7% of patients who underwent Mohs micrographic surgery consumed postoperative opioid pain medications but did not examine the risk of overprescription.⁵ We prospectively examined patients who were undergoing dermatologic surgery to determine (1) the incidence of opioid prescription after dermatologic surgery, (2) what percentage of opioid pain medications prescribed after dermatologic surgery are used in the postoperative period, and (3) patient and surgical characteristics associated with opioid pain medication prescription and use in an attempt to better understand postoperative opioid use. Understanding these factors is important to lay groundwork for evidence-based practices and to reduce the risk of opioid misuse and abuse.

Institutional review board approval was obtained at the University of Utah, Salt Lake City. A prospective observational pilot study was conducted to examine postoperative opioid pain medication use after dermatologic surgery among adult dermatologic surgery patients at the University of Utah in May and June 2011. All patients who were seen by the 4 dermatologic surgeons (K.D., G.B., M.H., and P.T.-F.) at the Department of Dermatology, University of Utah, were screened for and asked to participate in the study. The study was explained to the patients before the surgical procedure began, and written informed consent was obtained. Patients were included if they were 18 years or older, undergoing a single skin excision at a single surgical site (including Mohs micrographic surgery), able to consent to participate, and able to be reached by telephone on postoperative day 3 or 4. Patients who did not meet these criteria and those referred to another physician for further surgical treatment or repair were excluded. Study enrollment continued until 212 patients were included in the study (Figure 1).

Demographic and clinical information, including age, sex, surgical indication (basal cell carcinoma, squamous cell carcinoma, melanoma, and melanoma in situ, other), defect size in centimeters squared, repair method (granulation, primary closure, flap closure, and graft closure), and details of opioid prescription (opioid, dose, and dispensed number of pills), if prescribed, was collected on each participant. On postoperative day 3 or 4, the participants were called and asked to complete a short questionnaire about postoperative pain and medication use. All patients were asked about the use of over-the-counter medications, ice, other prescription pain medication, current and maximum pain on a scale of 0 to 10 (“0 is no pain and 10 is the worst pain ever experienced in your life”), and timing of maximum pain. Also, those who were prescribed opioids were asked whether they filled, and/or used pills from, the prescription; how many pills they used; how many pills they had left; and what they planned to do with any leftover pills. At the end of the telephone survey, patients with leftover pills were instructed regarding appropriate disposal. Because there were no significant differences in maximum pain, current pain, and percentage of opioid pills used between those called on postoperative day 3 and those called on postoperative day 4, the data were combined for all analyses.

Descriptive analyses of patient, surgical, and opioid prescription data were completed. Data from chart review and/or telephone questionnaires were used to determine the differences between (1) patients who were prescribed opioids and those who were not, (2) patients who used prescribed opioids and those who did not, and (3) patients who reported inadequate pain control and those who did not. To compare each of these groups, univariate analyses were performed using t tests and χ² or Fisher exact tests for continuous and categorical variables, respectively. Independent variables reaching significance (threshold, P < .05) were included in multivariate logistic regression models. Data from chart review and telephone questionnaires were used to model variables predicting maximum pain score using linear regression. Variables reaching significance (threshold, P < .05) were included in a multivariate linear regression model.

A total of 212 patients (age range, 34-94 years; mean [SD] age, 68 [13] years) completed the study (Figure 1). One hundred fifty-two of the participants (72%) were male. The most frequent reason for surgery was basal cell carcinoma (n = 141 [67%]) followed by squamous cell carcinoma (n = 56 [26%]), melanoma or melanoma in situ (n = 11 [5%]), and other (n = 4 [2%]). Defects ranged in size from 0.16 to 33.6 cm², with a mean (SD) defect size of 2.8 (3.6) cm². Primary closure (n = 92 [43%]) was the most frequent repair method followed by healing via granulation (n = 71 [34%]), flap closure (n = 40 [19%]), and graft closure (n = 9 [4%]).

Maximum pain scores ranged from 0 to 10, with a mean (SD) score of 3.5 (2.7). Most patients reported that their maximum pain occurred on the day of surgery (n = 109 [51%]) or on postoperative day 1 (n = 55 [26%]). Eighty-eight patients (42%) reported taking over-the-counter medications to help with pain, with acetaminophen being the one most frequently used (n = 66 [75%]) followed by nonsteroidal anti-inflammatory medications (n = 20 [23%]).

Opioid pain medication was prescribed to 72 of the 212 patients (34%). The number of opioid pills prescribed per patient ranged from 3 to 20, with a mean (SD) of 8.9 (2.7) pills. Patients who were prescribed opioids were younger, had larger defects, and were more likely to have undergone closure of the wound (primary, flap, or graft), as shown in Table 1. Defect size and closure type remained significant in multivariate analysis (Table 2).

Fifteen of the 72 patients (21%) who were prescribed opioids did not fill the prescription, and an additional 10 patients (14%) did not use any of the prescribed opioids; therefore, 25 patients (35%) did not use any of the prescribed opioid pills. The most frequent reason given for not filling or not using the opioid prescription was that the patient did not feel that an opioid was needed (11 of 15 [73%] and 10 of 10 [100%], respectively). Use of prescribed pills ranged from 0 to 100%, with a mean (SD) use of 32% (35%), which increased to 41% (34%) if the patients who did not fill the opioid prescription were excluded. Eight patients (11%) used all of the opioids prescribed. Looking only at the 57 patients who filled the prescription, the mean (SD) number of pills consumed per patient was 3.7 (3.7), and the mean (SD) number of leftover pills per patient was 5.4 (3.5). Forty-five of the 57 patients (79%) who filled an opioid prescription used 5 pills or fewer. A total of 301 pills were left unused in possession of the study participants. As shown in Figure 2, most of the patients with leftover pills (26 of 49 [53%]) planned to keep them, with only a minority (2 of 49 [4%]) planning appropriate disposal of the unused pills.

We examined the differences between the patients who used the opioids that were prescribed and those who did not (Table 3). Only maximum pain score was significantly different between those who used the opioid pain pills that were prescribed and those who did not. We attempted to further understand predictors of maximum pain. In univariate and multivariate analyses of all study participants, female sex, younger age, and graft closures were significantly associated with increasing maximum pain scores, as shown in Table 4; however, the overall fit of the model (multiple R² = 0.25) was limited, with much variation in maximum pain left unaccounted for in our model.

In addition to looking at opioid prescribing, we attempted to evaluate the patients who had an unmet need for pain control. Two patients (<1%) called their surgeon's office during the study period reporting a need for better pain control; both had not been prescribed an opioid but were prescribed one when they called. Fifteen of the 212 participants (7%) reported, when asked, that they needed better pain control. The only statistically significant patient characteristic discriminating those with inadequate pain control was closure type, with a higher frequency of postoperative pain with graft closures (Table 5).

In this prospective study, approximately one-third of patients (34%) were given prescriptions for opiods after dermatologic surgery. Most patients who were prescribed opioids were overprescribed, with only a fraction of prescribed opioids used (the mean number of pills prescribed per patient was 8.9 and the mean number of pills used per patient was 3.7, for an average use of 41% of prescribed opioids). Most patients (86%) who filled the opioid prescription had leftover pills, and most (53%) planned to keep the pills, creating an opioid reservoir for future misuse or abuse by the patient or anyone with access to the unused medications. Among the 4 surgeons in our study alone, more than 1800 unused opioid pills per year could be left over in patient homes for potential misuse or abuse. Furthermore, because dermatologic surgery is a common procedure, with more than 2 million skin cancers treated each year in the United States, opioid overprescription after dermatologic surgery may lead to millions of unused opioids nationally.⁶ Both reduced overprescription of opioids and patient instruction on appropriate disposal of unused pills may reduce the potential for diversion, misuse, and abuse.

While many patients were overprescribed opioids, a small minority (7%) reported unmet pain control need. Ideally, surgeons could avoid the overprescription and underprescription of opioids if they could accurately predict 2 parameters: which patients will require opioids and the correct quantity needed. Unfortunately, our data shed only minimal light on predicting which surgical cases will require opioid analgesia and at what dosage. Only increasing maximum pain score was associated with opioid use among those who were prescribed opioids. Flap and graft closures, however, were associated with increased maximum pain score, and patients with graft closures were more likely to report unmet pain control need, suggesting that patients with graft closures are more likely to require opioid analgesia. Firoz et al⁵ also found that patients who were undergoing full-thickness skin graft and flap closures had higher mean pain scores and greater narcotic prescription use than those undergoing a linear repair or secondary-intention healing after Mohs micrographic surgery. We are unsure why graft closures, independent of defect size, may result in greater pain, and this poses an area for future research.

There remains much room to better understand which patients will have postoperative pain requiring opioid analgesia and whether opioids are the most effective postoperative analgesics.⁷ Factors other than surgical characteristics (size, closure, and site) and simple patient demographics (age and sex) play a role in determining postoperative pain control needs. Cigarette smoking has been associated with pain syndromes, and anxiety and cognitive coping behaviors have been associated with increased postoperative pain.^8-10 Neurocognitive and behavioral factors, which were not measured in this study, may play a role and could be assessed preoperatively in the future.¹¹

Limitations to our study include the following: (1) the patients were located at a single institution and (2) there was a relatively short time frame (2 months) of data collection. Recall bias is an inherent limitation of this study because all information about pain and medication use was self-reported by patients and not independently verified. However, this study provides important new insight into opioid prescription after dermatologic surgery.

A minority of patients in our study required opioid analgesia after dermatologic surgery and, of those who did, most required 5 pills or fewer. Opioid overprescription occurs after dermatologic surgery, creating a reservoir of opioids that are available for misuse or abuse. Given the significant morbidity of opioid misuse and abuse, we advise cautious and limited use of opioids after dermatologic surgery. It is also important to instruct patients who are prescribed opioids about the appropriate disposal of unused pills.

Correspondence: KaLynne Harris, MD, Saltzer Medical Group, 9850 W St Luke's Dr, Ste 129, Nampa, ID 83687 (kalynne.harris@yahoo.com).

Accepted for Publication: September 22, 2012.

Published Online: December 17, 2012. doi:10.1001/jamadermatol.2013.1871

Author Contributions: Drs Harris and Tristani-Firouzi had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Harris, Bowen, Hadley, and Tristani-Firouzi. Acquisition of data: Harris, Curtis, Calder, Duffy, Bowen, and Hadley. Analysis and interpretation of data: Harris, Larsen, Bowen, and Hadley. Drafting of the manuscript: Harris, Curtis, Larsen, and Hadley. Critical revision of the manuscript for important intellectual content: Calder, Duffy, Bowen, Hadley, and Tristani-Firouzi. Statistical analysis: Harris. Obtained funding: Harris. Administrative, technical, and material support: Harris, Curtis, Larsen, Calder, Duffy, Hadley, and Tristani-Firouzi. Study supervision: Bowen, Hadley, and Tristani-Firouzi.

Funding/Support: The University of Utah Study Design and Biostatistics Center is funded in part by the Public Health Services research grants UL1-RR025764 and C06-RR11234 from the National Center for Research Resources.

Additional Contributions: Justin Endo, MD (Department of Dermatology, University of Wisconsin, Madison), helped with the study concept and design; Lisa Estrada (Department of Dermatology, University of Utah) helped with logistical support of the study; and Greg Stoddard, MS (University of Utah Study Design and Biostatistics Center), offered guidance with statistical analyses.

Additional Information: In memory of KaLynne Harris, MD.