We performed a retrospective, observational cohort study of patients treated for acne using electronic data from Health Data Compass, the enterprise data warehouse of the University of Colorado Anschutz Medical Campus and its affiliates, and from Children’s Hospital Colorado [10]. Data were obtained in the format of the Observational Medical Outcomes Partnership (OMOP) common data model from the Observational Health Data Sciences and Informatics (OHDSI) network. OHDSI is an interdisciplinary collaborative composed of an international network of researchers and health databases. The OMOP common data model was chosen for this study to facilitate future collaboration and expansion of this study to other sites within the OHDSI network and because of its methods library, which facilitates both large-scale implementation of observational study designs and large-scale data analytics. The cohort and outcome phenotype code and analytic code are available on GitHub [11].

Outcomes of interest were the type of oral antibiotic therapy prescribed for patients with acne diagnoses. All study subjects were aged between 10 and 45 years as of January 1, 2015, with a diagnosis of acne defined by at least 2 diagnoses of acne, as represented by select OHDSI concept IDs (Multimedia Appendix 1). While a variety of acne diagnoses fell into our inclusion criteria phenotype, only the following appeared in our population: acne, acne conglobata, acne varioliformis, excoriated acne, and tropical acne. Antibiotic start date was on or after January 1, 2015, occurring on or after any diagnosis of acne, with at least one year of follow-up from the first oral antibiotic prescription index date. Prescriptions for oral liquid or suspension forms were excluded. Similar data were collected regarding prescriptions for isotretinoin.

In an attempt to exclude antibiotics prescribed for reasons other than acne, the minimum antibiotic dose required was a prescription quantity of 28 or more, unless determined by the physician authors (MJA, TES) consensus to represent an equivalent quantity (eg, prescriptions for packs containing multiple doses).

We assessed the prevalence of antibiotics and isotretinoin prescriptions for the treatment of acne according to age and sex. Categorical values (sex, race, and ethnicity) were compared using chi-square tests, and continuous variables (age) were compared using 2-tailed t tests. All tests were evaluated at the P=.05 significance level. R (version 3.6.0; R Core Team) was used for all summaries and figures. We assessed whether the proportion of patients who received medication was different based on sex using an unadjusted logistic regression model with the binary outcome of receiving or not receiving medication.

This study is exempt from institutional review board assessment under category four of the University of Colorado Exempt Research Guidelines [12].

We characterized oral antibiotic use for acne treatment in a Colorado hospital system. Among 15,704 patients, we found that male patients were more likely than female patients to receive both antibiotic and isotretinoin prescriptions for acne. Among the eligible recipients of prescriptions for acne, there were 11,093 eligible female patients (70.6%) and 4605 eligible male patients (29.3%). However, among recipients of antibiotics, only 28% (3109/11,093) of female patients received prescriptions, compared to 39% (1810/4605) of male patients. Similarly, among recipients of isotretinoin prescriptions, only 10.4% (1159/11,093) of eligible female patients received prescriptions, compared to 21.5% (988/4605) of male patients. Male patients were 1.67 times as likely to receive antibiotics for acne and over twice as likely to receive isotretinoin for acne compared to female patients.

Previous research on the health burden of acne has revealed that women with acne demonstrate greater self-consciousness of their appearances and are more likely to seek care due to an increased subjective rating of severity [13]. Thus, female patients may seek care for less severe forms of acne compared to male patients, leading to a lower likelihood of being prescribed systemic medications for acne. Another potential explanation for this discrepancy could be that female patients have access to other systemic acne treatment options that male patients do not, namely, oral contraceptive pills (OCPs) and spironolactone [1,14-16]. This study did not include an analysis of antiandrogen medication.

Other studies suggest acne severity could be significantly higher in male patients, thus necessitating the need for systemic treatment [17]. However, further research is needed regarding the severity of acne presentations in male versus female patients, as we did not have this data and cannot say for certain that the female patients in our cohort had a less severe presentation compared to male patients.

The discrepancy in isotretinoin prescriptions may also reflect the increased burden of iPLEDGE, an FDA risk management program, on women capable of bearing children. In its initial introduction, iPLEDGE was associated with an initial 30% decrease in isotretinoin prescriptions for both men and women [18]. The system has requirements for patients, prescribers, pharmacies, and wholesalers. For women looking to start isotretinoin, iPLEDGE requires 2 negative pregnancy tests before starting therapy, monthly pregnancy tests during therapy, use of 2 forms of contraception during therapy, and use of 2 forms of contraception the month before and after therapy [19,20]. A 2013 pharmacy prescription claims–based study found that the overall number of isotretinoin prescriptions across all ages and sexes decreased after iPLEDGE implementation, with a greater decrease among women [19]. The extra requirements of iPLEDGE for women could be a dissuading factor, as they may be unable or unwilling to comply with the stringent contraception and testing requirements or may be considering pregnancy.

This study builds upon previous literature analyzing trends in antibiotic and isotretinoin prescriptions in dermatology (Table 5). In a large, retrospective study using OptumInsight data from 2004 to 2013, authors looked at prescribing patterns among dermatologists and nondermatologists for the treatment of acne [21]. They found a significant increase in spironolactone prescriptions for acne throughout that time period [21]. A separate retrospective OptumInsight study using data from 2007 to 2017 aimed to identify potential disparities in acne treatment and found that female patients were less likely than male patients to be prescribed both isotretinoin and oral antibiotics [6]. In addition, a retrospective population-based cohort study in British Columbia also found that individuals treated with isotretinoin were more likely to be male [5].

Minocycline was the most prescribed initial antibiotic and the most prevalent antibiotic in every year of data captured in this analysis (Figures 1 and 2). Beginning in 2013, doxycycline, the second most prevalent initial antibiotic, began a slow increase in prevalence, while minocycline prevalence began slowly declining. It is not completely clear why this trend occurred, as there is no overall clinical difference in efficacy between minocycline and doxycycline [22]. Side effects, such as an increased risk of lupus erythematous with minocycline, and patient and physician perceptions regarding the costs of the 2 drugs may have affected prescription patterns [22-25].

In contrast to this study, a cross-sectional analysis of antibiotic prescribing patterns among dermatologists from 2008 to 2016 found a decrease in antibiotic prescriptions for acne during that time [8]. We found no significant change in antibiotic prescription frequency in our cohort.

We used ICD codes, established by the Global Burden of Disease Study and validated by an international panel of skin disease experts, to determine the appropriate OHDSI acne concept IDs [26]. Some diagnoses may have been missed or inappropriately classified and were therefore not included in this study.

Similarly, some antibiotic prescriptions may have been missed, particularly in the years 2011-2014, where the total count of antibiotics for these data points was much lower than in the period 2015-2019. We used antibiotics cited in previous acne literature, and antibiotics not commonly prescribed for the treatment of acne or that are available over the counter were not captured [4,21]. Patients prescribed antibiotics for acne without an acne diagnosis would also have been missed, as would patients who were given a prescription for acne for a shorter time period than 28 days. We could also have potentially included patients who were on long-term antibiotics or isotretinoin for a reason other than acne, despite having 2 diagnoses of acne. Other data limitations include missing information on insurance, severity of disease, treatment duration, and prescription refills.

In our sample, male patients were 1.67 times more likely to have received an antibiotic prescription and over twice as likely to have received an isotretinoin prescription than were female patients. These findings could be attributed to the fact that female patients may seek treatment for less severe acne, resulting in fewer systemic therapy prescriptions. Further research is needed regarding the severity of acne presentations in male versus female patients, as we did not have this data and cannot say for certain that the female patients in our cohort had a less severe presentation compared to male patients. Side effect profiles and additional systemic treatment options for male patients (OCPs and spironolactone) may also play a role in this discrepancy between sexes.