Safety and Predictors of Clinically Significant Adverse Effects of Isotretinoin in Acne Vulgaris: A Retrospective Cohort Study from Saudi Arabia

Abstract

Purpose

Isotretinoin remains the most effective treatment for moderate-to-severe acne vulgaris; however, its use is frequently associated with mucocutaneous and metabolic adverse effects that may necessitate active clinical intervention. Data evaluating predictors of clinically significant toxicity in Saudi populations remain limited.

Patients and Methods

A retrospective cohort study was conducted at a tertiary academic hospital in Saudi Arabia. Electronic medical records of patients with acne vulgaris treated with isotretinoin between January 2020 and December 2024 were reviewed. Demographic variables, acne severity, isotretinoin dosing characteristics, laboratory findings, adverse effects, management strategies, and treatment outcomes were extracted. Multivariable regression analysis was performed to identify predictors of adverse effects requiring clinical intervention.

Results

A total of 444 patients were included. Most participants were aged 21–30 years (60.4%), and females represented 50.9% of the cohort. Clinical intervention was required in 70.7% of patients, including supportive care, treatment reduction and dose discontinuation to manage isotretinoin-related adverse effects. Mucocutaneous adverse effects, particularly cheilitis and xerosis, were the most frequently documented complications. Laboratory abnormalities were observed less frequently, although elevated LDL levels were identified in 9.9% of patients, followed by elevated total cholesterol (7.2%) and triglycerides (4.5%). Elevated liver enzymes were relatively uncommon (AST 4.1%, ALT 3.2%). Higher cumulative dose, dose escalation during therapy, and higher maximum dose significantly increased the likelihood of requiring clinical intervention. Female sex and higher body weight were associated with lower odds of intervention. Major improvement was achieved in 37.4%, complete clearance in 28.8%, giving a combined rate of 66.2% while recurrence after discontinuation occurred in 19.4%.

Conclusion

Isotretinoin, proved efficacious clinically in the management of acne vulgaris, however was identified to be associated with high number of adverse effects which require active management. Multivariable regression identified that the increasing dose and cumulative dose were the most influential determinants of the drug therapy. Particularly, lipid abnormalities (elevated LDL) were clinically relevant for the included cohort of patients. The outcomes therefore emphasize the need for a personalised approach to dosing along with continuous monitoring to achieve the optimal results.

Introduction

Acne vulgaris is a common chronic inflammatory disorder of the pilosebaceous unit, affecting approximately 9.4% of the global population and ranking among the most prevalent dermatological conditions worldwide, according to 2015 estimates. More recent data suggests the prevalence may vary according to region and time.¹ It predominantly affects adolescents, with up to 85% of individuals aged 12–24 years experiencing some degree of acne; however, a considerable proportion of patients continue to be affected into adulthood.^2,3 The condition is characterized by a multifactorial pathogenesis involving increased sebum production, abnormal follicular keratinization, colonization by Cutibacterium acnes, and an inflammatory response that leads to the formation of comedones, papules, pustules, nodules, and, in severe cases, permanent scarring.^4,5 Beyond its physical manifestations, acne has significant psychosocial consequences, including reduced quality of life, low self-esteem, anxiety, and depression, with a burden comparable to other chronic diseases such as asthma and epilepsy.^6,7

In Saudi Arabia, acne vulgaris represents one of the most frequently encountered dermatological conditions in outpatient settings, with reported prevalence rates ranging between 50% and 70% among adolescents and young adults.^8,9 The data represents epidemiological contrast regarding acne vulgaris, which affects about 9.4% of the global population. However, it has been estimated to affect up to 50–70% of adolescents and young adults in Saudi Arabia. The significant difference in prevalence rates for the same condition in different regions highlights the need for the analysis of regional factors that could be responsible for the huge burden of disease. Several regional factors may contribute to this high prevalence, including dietary patterns, environmental conditions, and hormonal influences. Female patients, in particular, may be more susceptible due to hormonal variations, including those associated with polycystic ovary syndrome, as well as the use of cosmetic products.⁸ Environmental factors such as temperature and humidity may also influence sebum production and exacerbate acne severity in certain regions of the country.¹⁰ Given the high burden of disease and its psychosocial impact, effective and safe treatment strategies are essential.

Isotretinoin, an oral retinoid, is considered the most effective treatment for moderate-to-severe acne vulgaris, particularly in cases resistant to conventional therapies such as topical agents and systemic antibiotics.^5,11 It acts through multiple mechanisms, including suppression of sebaceous gland activity, normalization of follicular keratinization, reduction of C. acnes colonization, and modulation of inflammation.⁵ Clinical indications for isotretinoin include severe nodulocystic acne, moderate acne unresponsive to standard treatments, acne associated with significant scarring, and conditions such as gram-negative folliculitis.¹¹ The drug demonstrates high efficacy, with long-term remission rates exceeding 80% following a single course of therapy.¹² The drug’s excellent efficacy is however always associated with a wide range of side effects that can occur in many body systems; therefore, the monitoring and clinical management of the drug are an essential component to its administration.

Despite its effectiveness, isotretinoin is associated with a broad spectrum of adverse effects that require careful monitoring. These include mucocutaneous effects, which are the most common and occur in the majority of patients, as well as systemic, metabolic, hematologic, musculoskeletal, neuropsychiatric, and reproductive adverse effects.^12,13 Notably, isotretinoin can significantly alter lipid metabolism, leading to hypertriglyceridemia and elevated cholesterol levels, and may also cause elevations in liver enzymes, necessitating regular laboratory monitoring. In addition, isotretinoin is a well-established teratogen, with a high risk of severe congenital malformations if exposure occurs during pregnancy.¹⁴ These risks have driven the global implementation of formal risk management policies. A number of countries require all patients taking isotretinoin to adhere to specific pregnancy prevention directives as well as regular laboratory monitoring. The SFDA (Saudi Food and Drugs Authority) regulates the use of isotretinoin in Saudi Arabia by requiring that all prescriptions follow established SFDA guidelines that include monthly pregnancy tests for all women who are capable of bearing children.¹⁵

Although the adverse effect profile of isotretinoin has been extensively studied in Western populations, there is comparatively limited evidence from Middle Eastern and specifically Saudi Arabian populations. Existing studies suggest potential regional differences in metabolic responses, including higher baseline rates of dyslipidaemia, Higher Baseline values of Dyslipidaemia (eg, with regional studies indicating a range of 20–30% in adult Saudis),¹⁶ which may Influence both frequency and severity of laboratory abnormalities induced by isotretinoin. Furthermore, genetic variability in enzymes involved in retinoid metabolism may contribute to inter-individual differences in drug response and toxicity.¹⁶ Environmental and lifestyle factors may also play a role in modifying treatment outcomes and adverse effect profiles in this population. The majority of the data on the adverse effects of isotretinoin have been collected from Western countries with very few studies focusing on middle eastern and particularly the Saudi Arabia population. Studies show that it appears that different regions may vary metabolically (dyslipidaemia) and therefore potentially affect the amount or type of laboratory abnormalities induced by isotretinoin. There are also a variety of dosing methods used in real-world settings such as the first dose selected, total dose target and titration schedule for isotretinoin that can be employed by physicians to manage their patients and that have not been evaluated in relation to the number of required clinical interventions. Understanding these patterns will provide insight into how doctors’ practices create safe environments for their patients during routine care.

A structured drug monitoring for isotretinoin treatment has become extremely important according to international and national guidelines. Which emphasizes the drug monitoring for better outcomes. The current understanding among experts recommends that patients undergo a baseline evaluation (ie, liver enzymes; fasting lipid profiles; pregnancy test) prior to starting isotretinoin treatment, which must be repeated every 1–2 months and also based on the patient’s specific clinical risk factors and progression.^17,18 Regular laboratory testing, is no longer widely recommended. Instead, guidelines recommend a risk-stratified assessment approach to limit excessive testing, yet ensure adequate patient safety.^19,20 Additionally, it is also strongly recommended that physicians adhere strictly to the structured measures for preventing pregnancy in females. Physicians should also educate patients about all possible mucocutaneous adverse effects due to taking isotretinoin, as well as document any psychiatric histories associated with each patient. All these updated recommendations clearly demonstrate the need for a balance between efficacy and safety while treating patients who have increased baseline metabolic risk.

Unlike previous published studies conducted in Saudi Arabia on side-effect of medications by Alajlan et al (2015),¹⁰ the current research has focused specifically on evaluating those factors that are predictive of clinically significant adverse events to better understand how they can be managed in a real-world clinical setting, focusing on specific predictive factors for clinically significant adverse drug reactions adds further importance to current study as we add an additional practice-based evidence.

Importantly, there remains a lack of comprehensive data on how adverse effects of isotretinoin are managed in clinical practice, particularly in terms of the frequency of clinical interventions, the types of management strategies employed, and their impact on treatment outcomes. Understanding these aspects is critical for optimizing patient care and informing evidence-based clinical guidelines. Therefore, this study aims to evaluate the safety profile, laboratory abnormalities, and clinical outcomes associated with isotretinoin therapy in a Saudi population, with particular emphasis on identifying predictors of adverse effects requiring clinical intervention and assessing real-world management practices.

Materials and Methods

Study Design

This study employed a retrospective observational cohort design using electronic medical records (EMRs) of patients who received isotretinoin therapy for acne vulgaris. The primary objective was to evaluate prescribing patterns, adverse effect profiles, laboratory abnormalities, and clinical outcomes associated with isotretinoin therapy over a five-year period. A retrospective design was chosen to capture routine clinical practice without influencing physician decision-making or patient management. Because no intervention was applied, all treatment decisions were made by the treating dermatologists in accordance with routine clinical protocols.

This is a retrospective study with a cross-sectional analysis was performed with in the cohort to assess the association between exposure and outcomes with each patient contributing data from a single course of isotretinoin therapy. This design allowed assessment of associations between exposure variables, including dosing strategies, demographic characteristics, and laboratory findings, and outcomes such as adverse effects and treatment response. The study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines,²¹ and all data extraction procedures were predefined to minimize bias and ensure consistency.

Study Setting

The study was conducted at the Dermatology Outpatient Clinic of King Khalid University Medical City (KKUMC), a tertiary-care academic medical center affiliated with King Khalid University in Abha, Saudi Arabia. The hospital serves approximately 2.2 million residents in the Asir region and provides specialized dermatological services, including a structured acne management program. The Asir region of Saudi Arabia, the study site of the current investigation, has a distinctive climate that features high humidity levels in combination with changing temperature values as a result of its topography. Thus, these climate factors can enhance oil (sebum) production, modify how well the skin functions as a barrier, and affect both the degree of severity of the acne and patient’s tolerance to treatment with isotretinoin.

Clinical care at KKUMC follows standardized protocols for prescribing isotretinoin and monitoring treatment outcomes, in accordance with Saudi Food and Drug Authority (SFDA) recommendations. The hospital uses a comprehensive EMR system that records patient demographics, diagnoses, prescribed medications, laboratory results, follow-up visits, and clinical outcomes. This robust digital infrastructure enabled accurate, comprehensive data collection for the present study.

Study Period

Data were collected from all eligible patients who received isotretinoin therapy from 1 January 2020 to 31 December 2024. This five-year period was selected to ensure an adequate sample size and to reflect contemporary clinical practices at the institution. COVID-19 (2020–2021) most likely had an impact on follow up frequency, patient’s drug adherence and thorough reported any adverse reactions from treatments, these impacts were recognized as limitations.

Study Population and Sample Size Calculation

Eligible participants were patients with a confirmed diagnosis of acne vulgaris (ICD-10 code L70.0) made by a board-certified dermatologist at KKUMC. Patients were required to be at least 16 years of age at the initiation of isotretinoin therapy and to have completed at least 30 days of treatment. To ensure the reliability of clinical and laboratory assessments, only patients with documented baseline laboratory investigations within 30 days before treatment initiation were included. These investigations comprised a complete blood count (CBC), a fasting lipid profile including low-density lipoprotein (LDL), high-density lipoprotein (HDL), total cholesterol, and triglycerides, and liver function tests, specifically aspartate aminotransferase (AST) and alanine aminotransferase (ALT). Furthermore, patients were required to have documented baseline acne severity and at least one follow-up clinical assessment.

Patients were excluded if their medical records were incomplete or lacked critical data needed for analysis. Individuals with other chronic dermatological conditions, such as psoriasis, rosacea, or lupus erythematosus, were excluded to prevent confounding of clinical or laboratory outcomes. In addition, patients receiving systemic medications known to interact with isotretinoin or to independently cause similar adverse effects were excluded such as tetracyclines, vitamin A derivatives and anticonvulsants which cause similar adverse drug reactions, unless these factors were clearly documented and accounted for in the analysis.

The sample size was calculated using the formula for a single-proportion estimate, where Z represents the z-score for the desired confidence level (eg, Z = 1.96 for 95% confidence). Based on a retrospective study by Rademaker et al (2010), approximately 47% (p = 0.47) of patients receiving isotretinoin experience at least one adverse effect requiring clinical intervention. With a 5% margin of error (d = 0.05), the minimum required sample size was 384 patients.¹² As no Saudi or regional estimates were available. So, the western figures were used as conservative baseline for the final sample size calculation ensuring adequate power. As this retrospective study included data collected over 5 years at a tertiary-care center, the available sample was expected to exceed this minimum, enhancing statistical power for subgroup analyses (eg, sex, acne severity, and dosing). To account for missing laboratory data, a 10% contingency was added, yielding a target sample size of 422 patients. The final sample included 444 patients, providing sufficient power to analyze both primary and secondary outcomes.

Data Collection

Data were extracted from EMRs using a structured, standardized data collection form developed before study initiation. Collected variables included demographic characteristics (age, sex, and body weight) and clinical variables (baseline acne severity and relevant comorbidities, when available). Treatment-related variables were also recorded, including the initial isotretinoin dose, any dose adjustments during therapy, cumulative dose, and treatment duration.

Laboratory data were systematically collected, focusing on lipid parameters—LDL, HDL, total cholesterol, and triglycerides—as well as liver function indicators (AST and ALT) and hematological parameters from the CBC. Adverse effects were documented and categorized into clinically relevant groups, including mucocutaneous, metabolic or laboratory-related, musculoskeletal, neuropsychiatric, and other reported effects.

In addition, data on clinical management were collected, including whether patients required clinical intervention to manage adverse effects. Such interventions included dose reduction, temporary discontinuation of therapy, initiation of adjunctive treatments, or patient counseling. Treatment outcomes were also recorded, including degree of clinical improvement (minimal, moderate, or major), recurrence after treatment discontinuation, and overall treatment completion.

Variables and Operational Definitions

The primary outcome variable was the need for clinical intervention to manage adverse effects associated with isotretinoin use (binary: yes/no). Secondary outcome variables included (1) types of documented adverse effects, (2) laboratory parameter abnormalities, (3) post-treatment acne severity classification, and (4) acne recurrence following treatment discontinuation. Key independent predictor variables included patient demographics, anthropometric measures, acne severity, dosing parameters, concomitant medications, and trigger factors. All variables were operationally defined in the data extraction form. Acne post-treatment severity assessment was based on dermatologist documentation during the last visit with patients for the purpose of assessing a severity level in accordance with those used at baseline (nodulocystic, mild, moderate, severe) as well as recurrence of inflammatory lesions that prompted a documented restart of systemic medication or an escalation of treatment within 6 months after discontinuing isotretinoin.

A clinical intervention is referred to any action that the physician took in response to an adverse effect related to isotretinoin. Examples of interventions were pharmacologic actions (reduction of dose, temporary cessation of therapy, addition of other medications) and formal non-pharmacologic actions are well explained as patient counselling, which was documented as advising on lifestyle modifications. However, simply providing assurance was not considered to be a clinical intervention.

The burden of side effects was defined as number of side effects experienced by each patient during the course of treatment. For regression analysis, the variables are categorised in three groups, low- 0-1, moderate 2–3 and high ≥ 4 adverse drug reactions. The method was chosen to avoid misinterpretation and to reflect logic that greater side effects increase the likelihood of intervention.

For laboratory values, “abnormal” was defined as any value falling outside the reference range established by the laboratory at King Khalid University Medical City, the specific upper limits of normal (ULN) used for AST, ALT, total cholesterol, LDL, and triglycerides are provided in Supplementary Table 1. Laboratory abnormalities were determined by hospital reference range values. Since grade levels from the Common Toxicity Criteria for Adverse Events (CTCAE) were not used to define laboratory abnormalities in this study, mildly elevated laboratory results do not necessarily represent a clinically relevant or toxic condition. Body mass index (BMI) was categorized according to World Health Organization (WHO) criteria as follows: underweight (<18.5 kg/m²) only, normal weight (18.5–24.9 kg/m²), overweight (25.0–29.9 kg/m²), and obese (≥30.0 kg/m²). Patients with incomplete records were excluded at the baseline. Therefore, the final data set contains no missing values for included variables.

Data Management and Statistical Analysis

All extracted data were entered into a secure database and reviewed for completeness and accuracy prior to analysis. Data cleaning included consistency checks and verification of outliers. Descriptive statistics were used to summarize patient demographics, clinical characteristics, treatment patterns, laboratory findings, and adverse effects. Continuous variables were presented as means with standard deviations or medians with interquartile ranges, depending on data distribution, while categorical variables were reported as frequencies and percentages.

Inferential analyses were conducted to evaluate associations between acne severity and clinical variables. Chi-square tests were used for categorical comparisons, while independent t-tests or nonparametric equivalents were applied for continuous variables as appropriate. Pearson correlation coefficients (r) were calculated to estimate the strength and direction of associations.

Multivariant logistic regression model was selected based on an assessment towards the cause of adverse drug reactions that necessitate medical intervention. The selected model allows multiple variables to be tested at ones, a variable of interest “need for clinical intervention” was treated as 1 = need for clinical intervention and 0 = no need for clinical intervention. Variables were separated in to groups based on categories such as demographics, anthropometric measures, the acne severity, acne triggering factors, aspects of the treatment regimen such as initial dose in milligrams per day, cumulative dose in milligrams, dose increase during the course of treatment, highest dose level reached by the patient, and abnormal blood values.

Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported. Model fit was assessed using the likelihood ratio chi-square test and pseudo-R² statistics (Cox–Snell and Nagelkerke). Multicollinearity was evaluated prior to model interpretation. Variance inflation factors (VIFs) were all <2, indicating no significant multicollinearity among predictor variables. A p-value < 0.05 was considered statistically significant. All analyses were performed using IBM SPSS Statistics for Windows, version 29.0 (IBM Corp., Armonk, NY, USA).

Omnibus model fit for each adverse effect category was tested by means of a likelihood ratio Chi-square-test. The models that showed a good fit (p < 0.05) were judged to be valid. Predictors from models showing no good fit (ie, p > 0.05), although they were reported in descriptive terms, are therefore not treated as reliable.

Results

Participants’ Demographics

Data from all 444 participants with acne vulgaris treated with isotretinoin were analyzed, with no missing values reported (Table 1). The majority of participants were aged 21–30 years (60.4%), followed by adolescents aged ≤ 20 years (33.9%), while only a small proportion were aged above 30 years. This distribution is consistent with the epidemiology of acne, which predominantly affects adolescents and young adults. The gender distribution was nearly equal, with females comprising 50.9% and males 49.1% of the sample.

Table 1.

Participant Demographics and Baseline Characteristics (N = 444)

Characteristic	N	%
Age group (years)
11–20	146	32.9%
21–30	268	60.4%
31–40	20	4.5%
41–50	10	2.3%
Sex
Male	218	49.1%
Female	226	50.9%
Body weight (kg)
31–50	74	16.7%
51–70	188	42.3%
71–90	134	30.2%
91–110	36	8.1%
111–130	8	1.8%
131–140	4	0.9%
Height (m)
1.30–1.49	20	4.5%
1.50–1.59	112	25.2%
1.60–1.69	172	38.7%
1.70–1.79	92	20.7%
1.80–1.89	40	9.0%
1.90–1.99	8	1.8%
Body Mass Index (BMI, kg/m2)
Underweight (< 18.5)	76	17.1%
Normal weight (18.5–24.9)	294	66.2%
Overweight (25–34.9)	58	13.1%
Obese (≥35)	16	3.6%

Note: Values are expressed as frequency (n) and percentage (%).

Abbreviation: BMI, body mass index.

Anthropometric characteristics indicated that most participants had body weights between 51–70 kg (42.3%) and 71–90 kg (30.2%). The most common height range was 1.60–1.69 m (38.7%), followed by 1.50–1.59 m (25.2%). Based on these measures, the majority of participants had a normal body mass index (BMI), while smaller proportions were underweight, overweight, or obese. Overall, the demographic profile reflects a typical acne population predominantly composed of young adults with balanced gender representation and generally normal anthropometric status.

Clinical Characteristics

At baseline, most participants presented with moderate (50.0%) or severe acne (38.3%), while only a small proportion had mild or nodulocystic acne (Table 2 and Figure 1). Overall, more than 90% of patients had moderate-to-severe disease, indicating that the majority were appropriate candidates for systemic isotretinoin therapy. Hormonal changes during puberty were the most commonly reported trigger factor, followed by combinations of hormonal, dietary, and genetic influences. Additional triggers included excess sebum production, menstrual-related hormonal fluctuations, and lifestyle factors, although medication-induced acne was relatively uncommon (3.6%).

Table 2.

Clinical Characteristics of Acne Vulgaris Patients Treated with Isotretinoin (N = 444)

Characteristic	N	%
Acne severity
Mild	18	4.1
Moderate	222	50.0
Severe	170	38.3
Nodulocystic	34	7.7
Total duration of isotretinoin therapy
0–4 months	116	26.1
5–8 months	222	50.0
9–12 months	98	22.1
13–16 months	8	1.8
Initial daily dose (mg)
0–10 mg	18	4.1
11–20 mg	128	28.8
21–30 mg	164	36.9
31–40 mg	132	29.7
41–50 mg	2	0.5
Cumulative dose (mg)
Less than 120 mg/kg	80	18.0
120–150 mg/kg (therapeutic range)	338	76.1
>150–225 mg/kg	24	5.4
>225 mg/kg	2	0.5
Dose escalated during therapy
Yes	100	22.5
No	344	77.5
Concomitant acne medications prescribed
Yes	116	26.1
No	328	73.9
Intervention required for side effect management
Yes	314	70.7
No	130	29.3
Action taken for side effect management
Isotretinoin discontinued	24	5.4
Isotretinoin dose reduced	6	1.4
Added another medication	78	17.6
Non-pharmacological advice only	64	14.4
No intervention needed	114	25.7
Medication added + non-pharmacological advice	106	23.9
Dose reduced + medication + non-pharmacological advice	52	11.7
Acne status after completing isotretinoin
Complete clearance	128	28.8
Major improvement	166	37.4
Minimal or no improvement	150	33.8
Acne recurrence after stopping isotretinoin
Yes	86	19.4
No	232	52.3
Not documented	126	28.4
Follow-up visits during therapy
1–2 visits	102	23.0
3–4 visits	252	56.8
5–6 visits	88	19.8
7–8 visits	2	0.5
Documented non-compliance
Yes	120	27.0
No	324	73.0
Laboratory abnormalities
ALT abnormal	14	3.2
AST abnormal	18	4.1
Total cholesterol abnormal	32	7.2
LDL abnormal	44	9.9
Triglycerides abnormal	20	4.5
CBC abnormal	14	3.2
Serum urea abnormal	2	0.5
Serum creatinine abnormal	2	0.5

Note: Values are expressed as frequency (n) and percentage (%).

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; LDL, low-density lipoprotein; CBC, complete blood count.

Figure 1 .

Percentage distribution of common adverse effects among patients treated with isotretinoin.

Regarding treatment characteristics, isotretinoin therapy most frequently lasted between 5–8 months (50.0%), followed by 9–12 months (22.1%) and less than 4 months (26.1%). Initial daily dosing was most commonly within the range of 21–30 mg (36.9%) and 31–40 mg (29.7%), while the majority of patients (76.1%) achieved a cumulative dose within the recommended therapeutic range of 120–150 mg/kg. Dose escalation was required in 22.5% of patients, suggesting the need for individualized dose adjustments during therapy. Concomitant acne treatments were prescribed in 26.1% of patients, most commonly topical clindamycin, followed by benzoyl peroxide and salicylic acid preparations.

Clinical intervention to manage adverse effects was required in 70.7% of patients (primary outcome), highlighting the substantial burden of treatment-related side effects in routine practice. Management strategies most commonly included a combination of pharmacological and non-pharmacological approaches, while some patients required dose reduction or treatment discontinuation. The most frequent clinical scenarios prompting intervention involved combinations of mucocutaneous symptoms (eg, cheilitis and dry skin) with other manifestations such as nosebleeds or ocular dryness. Dose escalation during therapy was strongly associated with the need for clinical intervention (OR = 93.714, 95% CI: 11.873–739.691, p < 0.001) (Secondary outcomes).

Laboratory abnormalities were relatively infrequent but clinically relevant. Elevated low-density lipoprotein (LDL) levels were the most common abnormality (9.9%, (n = 44)), followed by increased total cholesterol (7.2%, (n = 32)) and triglycerides (4.5%, (n = 20)). Mild elevations in liver enzymes were observed in a small proportion of patients (AST 4.1% (n = 18), ALT 3.2%, (n = 14)), while renal abnormalities were rare. These findings indicate that lipid disturbances were more prominent than hepatic or renal abnormalities in this cohort.

In terms of treatment outcomes, 66.2% (combined cohort of major improvement and complete clearance) of patients achieved at least moderate improvement, with 37.4% demonstrating major improvement and 28.8% achieving complete clearance. Acne recurrence was evaluable in 318 patients (71.6%); among these, despite the high effectiveness of isotretinoin, acne recurrence was observed in 19.4% of patients after treatment discontinuation, whereas 52.3% experienced no recurrence. The remaining 28.4% had undocumented recurrence status and were excluded from the primary recurrence analysis. A sensitivity analysis coding missing recurrence as “unknown” yielded similar conclusions. Follow-up patterns showed that most patients attended 3–4 clinical visits during therapy, indicating moderate engagement with healthcare services. Medication adherence was generally high, with 73.0% of patients demonstrating full compliance, although a notable proportion (27%) exhibited some degree of non-adherence.

Association Between Acne Severity and Clinical Variables

Chi-square tests were conducted to assess associations between acne severity (mild, moderate, severe, nodulocystic) and demographic and clinical variables among 444 patients treated with isotretinoin (Table 3). Statistical significance was set at p ≤ 0.05, and Effect sizes were quantified using Cramer’s V values.

Table 3.

Chi-Square Analysis of Associations Between Acne Severity and Patient Characteristics (N = 444)

Variable	Mild (n=18)	Moderate (n=222)	Severe (n=170)	Nodulocystic (n=34)	χ2	df	p-value	Cramer’s V
Age group (years)	6	64	92	8	12.944	9	0.165	0.099
Sex (M/F)	04 / 14	100 / 122	92 / 78	22 / 12	11.69	3	0.009	0.162
Body weight (kg)	6	44	20	4	22.911	15	0.086	0.131
Height (m)	2	8	0	0	102.826	18	< 0.001	0.278
BMI (kg/m2)	2	44	24	6	12.954	9	0.165	0.099
Need for intervention	4	56	144	26	52.384	3	< 0.001	0.343
Side effect burden	2	12	30	4	165.87	63	< 0.001	0.353

Note: Bold p-values indicate statistical significance (p <0.05). Variables Include Age Group, Sex, Body Weight, Height, BMI, Need for Intervention, and Side Effect Burden. Values Represent Patient Counts Distributed Across Severity Categories.

Abbreviation: BMI, body mass index.

A significant association was observed between acne severity and gender (χ² = 11.69, p = 0.009, Cramer’s V = 0.162), with males more frequently represented in severe and nodulocystic cases, while females were more likely to have mild or moderate acne. Height was also significantly associated with severity (χ² = 102.83, p < 0.001, Cramer’s V = 0.278), indicating a moderate positive relationship, whereas age, weight, and BMI showed no significant associations.

Clinical severity was strongly associated with the need for intervention due to adverse effects (p < 0.001, Cramer’s V = 0.343), with higher severity linked to greater intervention requirements. The burden of adverse effects requiring intervention showed the strongest association with severity (p < 0.0001, Cramer’s V = 0.353). Treatment outcomes were also associated with baseline severity (p < 0.01), with better outcomes in mild cases and poorer outcomes in severe and nodulocystic acne. Recurrence was more common in severe forms (p < 0.002).

No significant association was found between acne severity and number of follow-up visits. However, documented non-compliance (non-adherence) was significantly associated with severity (p < 0.02). Overall, the strongest relationships were observed for adverse effect burden, height, and need for clinical intervention.

Predictors of Specific Adverse Effects

Multiple logistic regression models were developed to identify predictors of specific adverse effects (Table 4). Several models failed the omnibus tests (Dry skin, Nosebleeds, Lab abnormalities, Dry eyes, Others; p > 0.05) indicating that those are not statistically significant. Even though many of the Individual Predictors had statistically significant P-values, must be considered explanatory rather than causal findings. Because all of the overall models for dry skin, laboratory abnormalities, dry eyes, nosebleeds and “other” adverse effects were not statistically significant, individual coefficient estimates from these models are presented only for descriptive purposes and should not be interpreted as independent predictors.

Table 4.

Model Fit Statistics for Logistic Regression Models Predicting Adverse Effects (N = 444)

Outcome (Model)	DV	Omnibus χ2	df	p	–2 Log L	Cox–Snell R2	Nagelkerke R2	Accuracy	Model fit	Significant predictors
Need for intervention	Need for intervention	331.602	17	< 0.001	205.312	0.526	0.75	90.50%	Significant	Cumulative dose / Dose escalation / Maximum dose
Dry lips/cheilitis	Dry Lips	182.541	47	< 0.001	119.653	0.567	0.756	–	Significant
Dry skin	Dry Skin	52.583	41	0.106	248.73	0.214	0.286	–	Not significant
Nose bleeds	Nose bleeds	53.141	41	0.097	154.673	0.216	0.352	–	Not significant
Lab abnormalities	Lab Abnormalities	35.678	41	0.706	169.12	0.151	0.248	–	Not significant
Dry eyes	Dry Eyes	52.872	41	0.101	89.741	0.215	0.449	–	Not significant
Others	Others	34.632	41	0.748	157.461	0.147	0.251	–	Not significant

Notes: Accuracy = classification accuracy. Significant = omnibus p <0.05. Models with non-significant omnibus tests (p > 0.05) are shown for completeness; individual predictors from these models should not be regarded as statistically validated.

Abbreviations: DV, dependent variable; –2 Log L, –2 log likelihood; Cox–Snell R² and Nagelkerke R², pseudo-R² statistics;

Dry lips/cheilitis were significantly associated with female sex, management strategies, and LDL abnormalities. Dry skin was associated with height and management strategies. Nosebleeds were linked to body weight, height, LDL abnormalities, and the number of side effects requiring intervention.

Laboratory abnormalities were predicted by management strategies and LDL levels. Dry eyes were associated with management strategies, LDL abnormalities, and side-effect burden. Other adverse effects (eg, menstrual irregularities, joint pain, headache) were significantly associated with longer treatment duration, higher initial doses, management strategies, and LDL abnormalities.

Overall, LDL abnormalities and treatment-related factors (dose and duration) consistently emerged as important predictors across multiple adverse effect categories.

Determinants of Clinical Intervention (Multivariable Logistic Regression)

Multivariable logistic regression analysis identified potential predictors for clinical interventions resulting from adverse effects of isotretinoin as shown in Table 5. Odds ratios indicate that female subjects (p = 0.009; OR = 0.21), subjects with higher body weights (p = 0.045; OR = 0.32), and higher numbers of acute or chronic acne triggers (p = 0.005; OR = 0.93) had lower odds of an intervention. Higher initial doses of isotretinoin are negatively correlated with an intervention (OR = 0.231; p < 0.001), while higher total cumulative dose (OR = 3.75; p = 0.021), dose escalations (OR = 92.71; p < 0.001), and higher maximal doses (OR = 1.09; p = 0.003) all have positive associations with the need for an intervention. Use of concomitantly prescribed medications has a protective association (OR = 0.49; p = 0.006). Higher burden of side effects was significantly associated with higher odds of requiring intervention (p < 0.001), consistent with clinical expectations. Cases with moderate or high burden (≥ 2 adverse drug effects) were more likely to require clinical management of side effects compared to those with low burden.

Table 5.

Multivariable Logistic Regression Predictors of Clinical Intervention for Isotretinoin-Related Adverse Effects (N = 444)

Variable	B	SE	Wald	p	OR	95% CI
Age group	0.511	0.328	2.429	0.119	1.667	0.877–3.171
Sex (female vs male)	–1.581	0.606	6.795	0.009	0.206	0.063–0.676
Body weight	–1.135	0.566	4.023	0.045	0.321	0.106–0.974
Height	–0.322	0.304	1.12	0.29	0.725	0.400–1.315
BMI	–0.739	0.618	1.427	0.232	0.478	0.142–1.605
Acne severity	0.014	0.376	0.001	0.971	1.014	0.485–2.118
Trigger factors	–0.074	0.026	7.928	0.005	0.928	0.881–0.978
Medication-induced acne	–0.308	0.342	0.811	0.368	0.735	0.376–1.437
Duration of therapy	–0.271	0.262	1.069	0.301	0.763	0.456–1.275
Initial daily dose	–1.466	0.329	19.823	< 0.001	0.231	0.121–0.440
Cumulative dose	1.322	0.573	5.324	0.021	3.752	1.220–11.534
Dose escalated during therapy	4.54	1.054	18.552	< 0.001	93.714	11.873–739.691
Maximum dose reached	0.087	0.029	8.867	0.003	1.091	1.030–1.156
Other acne medications prescribed	–1.364	0.773	3.112	0.078	0.256	0.056–1.164
Concomitant medication type	–0.715	0.26	7.536	0.006	0.489	0.294–0.815
Side effect burden (moderate/high vs low)	–0.260	0.035	55.227	< 0.001	0.771	0.720–0.826

Note: Side effect burden was categorized as low (0–1 adverse effects), moderate (2–3), and high (≥4). Reference category: low burden. B = regression coefficient. Dependent variable coding: 1 = intervention required, 0 = no intervention. Bold p values indicate statistical significance (p <0.05). Omnibus χ²(16) = 331.602, p <0.001; Nagelkerke R² = 0.750; Overall classification accuracy = 90.5%.

Abbreviations: SE, standard error; Wald, Wald chi-square statistic; OR, odds ratio [= Exp(B)]; CI, confidence interval.

Discussion

This study provides comprehensive evidence from routine clinical practice on the safety profile, laboratory abnormalities, and clinical outcomes associated with isotretinoin therapy in a Saudi population. The findings demonstrate that while isotretinoin is highly effective, its use is associated with a substantial burden of adverse effects requiring active clinical management, with 70.7% of patients requiring intervention. Interventions varied from supportive actions (such as moisturizers and counseling about hydration) to pharmacological actions like dose reduction or temporary stopping. Most interventions were supportive and fewer required dose changes. This proportion is higher than that reported in some previous regional and international studies, where intervention rates ranged from approximately 40% to 50%.²² This discrepancy may reflect differences in environmental factors, patient characteristics, and clinical practice patterns, particularly in regions with harsher climates such as the Asir region, which remains as a hypothesis which requires geographical and dermatological validation. Similar trends have also been reported in Saudi populations.²³

Mucocutaneous adverse effects, particularly cheilitis and xerosis, were the most frequently reported, consistent with the pharmacological mechanism of isotretinoin in suppressing sebaceous gland activity.²⁴ Previous studies have reported cheilitis in over 90% of patients, making it a near-universal adverse effect and a marker of drug activity.²⁴ Comparable findings have been documented in Saudi Arabia, where dryness-related symptoms were the most common adverse effects among isotretinoin users.²³ However, the clinical significance of these adverse effects lies not in their frequency but in their impact on patient adherence and quality of life, which often necessitates clinical intervention.

A key finding of this study is the prominent role of lipid abnormalities, particularly LDL elevation, as indicators of systemic toxicity. Elevated LDL was observed in 9.9% of patients and was significantly associated with multiple adverse effects across different models. This aligns with previous studies demonstrating that isotretinoin disrupts lipid metabolism, leading to hyperlipidaemia in a dose-dependent manner.^5,25 Consistent with this, studies conducted in Saudi Arabia have reported significant increases in lipid parameters during isotretinoin therapy.¹⁶ Interestingly, liver enzyme elevations were relatively infrequent compared to historical reports (typically 10–15%),^26,27 suggesting that lipid parameters may serve as more sensitive and clinically relevant markers of toxicity in this population. These results indicate lipid parameters such as LDL might provide clinically useful indicators of toxicity in the included cohort of Saudi Arabia; however, LDL should be considered with regard to additional laboratory measures for clinical interpretation, and future prospective studies are required to determine if LDL is an appropriate first-line marker to monitor toxicity with isotretinoin therapy. Most abnormal lipid levels were lab findings and not symptomatic events. Also, when liver enzymes do rise they may be of greater clinical significance than most mild lipid changes.

Another important contribution of this study is the identification of dose-related factors as key predictors of adverse effects. The higher cumulative dose, the higher dose escalation, and the higher maximum dose contributed to an increase in the chance that an individual needs intervention. This is similar to what would occur with a dose–response relationship. However, because this study was conducted retrospectively, no causal link can be made with these findings as such and need for a prospective study is imperative to evaluate the isotretinoin toxicity.²⁸ The association between higher initial dose and increased risk of adverse effects highlights the importance of cautious dose initiation and gradual titration. These findings reinforce current clinical recommendations advocating individualized dosing strategies rather than fixed high-dose regimens.

The toxic side effects from isotretinoin are directly related to the dose of isotretinoin. However, the current study identified a correlation with no significant difference in having to stop treatment due to adverse reactions among those that received the maximum first day’s dose versus those who did not receive the maximum dose. This could be due to the physician’s choice of selecting least risky patients with the high dose of isotretinoin. On the other hand, the other measures such as the patient’s cumulative dose over time, the frequency of dosing and the dose increased during treatment, and the maximum dose of isotretinoin the patient was prescribed, demonstrated a statistically significant association with the need for intervention, which clearly highlights the dose – adverse effect linear relationship. The regression analysis post recoding the variables in to categorical values, the model showed the increasing side effects burden increases the chance for an clinical intervention. It must be noted that several adverse effects failed to reach a statistical significance, although there was a clear association with each predictor individually. The inability of overall model fit indicates a limitation associated with using retrospective data which can lead to inadequately identified risk factors. Therefore, these results are exploratory in nature and require further research through prospective study design.

The study also identified demographic and anthropometric factors that influence the risk of adverse effects. Female sex was associated with a lower likelihood of developing certain adverse effects, such as cheilitis, while higher body weight and height were associated with reduced risk of nosebleeds. These findings may reflect differences in pharmacokinetics, body composition, or behavioral factors, including adherence to supportive measures such as hydration and skincare. Although these associations have not been consistently reported in previous studies, they highlight the need for further research into patient-specific predictors of isotretinoin tolerability. The potential association between these characteristics could be due to a diverse reason; pharmacokinetic parameters, physical characteristics or even possibly behaviour. However, due to the current study being retrospective with low levels of generalization, caution is recommended while interpreting the results. Additionally, validation of the results through a larger prospective study would provide stronger support for any association that exists.

From a clinical perspective, one of the most important findings is that despite the high prevalence of adverse effects, treatment discontinuation was rare (< 5%), indicating that most adverse effects are manageable with appropriate intervention. This underscores the importance of proactive clinical management, including dose adjustments, supportive therapies, and patient education. These findings are consistent with previous studies demonstrating that isotretinoin therapy can be safely continued in most patients when adverse effects are appropriately managed.²³

In terms of effectiveness, isotretinoin demonstrated a high rate of clinical improvement, with 66.2% of patients achieving moderate to major improvement, consistent with previously reported remission rates exceeding 80%.¹² However, the observed recurrence rate of 19.4% highlights the need for long-term follow-up and potential maintenance strategies, particularly in patients with severe acne. Similar recurrence patterns have also been reported in regional studies, including those conducted in Saudi Arabia.¹¹ Previous studies have suggested that recurrence is associated with lower cumulative doses and more severe baseline disease,^29–31 which may explain the findings observed in this study. The extremely high odds ratio shows that adjusting an individual’s isotretinoin dosage during their treatment regimen is a significant risk factor for adverse events necessitating intervention. The wide confidence interval (CI) in this finding suggests there may be considerable variation; however, it does suggest the necessity to cautiously initiate dosing as well as minimize increases throughout the therapy (Table 5).

The findings of this study have several important implications for clinical practice and healthcare policy. First, clinicians should adopt individualized dosing strategies, with careful consideration of patient characteristics and gradual dose escalation to minimize adverse effects. Second, routine monitoring protocols should prioritize lipid profiles, particularly LDL, as key indicators of systemic toxicity, potentially more informative than liver function tests in this population. The most significant scientific contribution of the current research comparing to the literature globally, is its details regarding clinical intervention and lipid abnormalities within the context of a Saudi population. The data can be used as regional specific information to help guide and monitor different approaches for managing the clinical issues. Based on findings like these lipid profiles should be monitored right at the start and every two to 3 months during treatment. If LDL levels go above 160 mg/dL or triglycerides go above 300 mg/dL then closer evaluation should happen and dose adjustment might be considered.

From a policy perspective, these findings support strengthening pharmacovigilance systems for isotretinoin in Saudi Arabia. Standardized monitoring protocols, improved documentation of adverse effects, and integration of electronic alerts within EMRs could enhance patient safety. Additionally, national guidelines may benefit from incorporating region-specific recommendations based on local patient characteristics and environmental factors.

Improving patient awareness is essential to optimize treatment outcomes. Patients should be counseled on the expected adverse effects of isotretinoin, the importance of adherence to treatment and monitoring schedules, and strategies to mitigate side effects, such as the use of moisturizers and adequate hydration. Enhanced patient education may reduce anxiety, improve adherence, and minimize unnecessary discontinuation of therapy.

This study has several strengths, including a relatively large sample size, the use of real-world clinical data, and comprehensive assessment of both clinical and laboratory outcomes. The inclusion of detailed dosing variables and analysis of predictors of adverse effects provides valuable insights into clinical practice.

However, several limitations should be acknowledged. The retrospective design limits the ability to establish causal relationships and may be subject to documentation bias. The retrospective study design of the current study and missing data of recurrence of acne is a limitation to making causal inferences and may have an issue with documenting the data. Additionally, a single centered design can limit the generalizability of the results because the type of patients treated at a centre or the practice styles employed by centres may vary from one region to another. The above limitations on the study are important factors for understanding the results. Hence, these limitations should be taken into consideration when analysing and interpreting the given results. Although there was a measure of compliance (documented non-compliance: 26.6%) from patients, it could not be used as an input for the regression analysis because of model limitations. Furthermore, While documented nonadherence (26.6%) may have had an influence from adverse effects, there were other factors that could have led to a contribution in follow-up disruptions caused by the COVID-19 Pandemic.

Future research should focus on prospective, multicenter studies to validate these findings and enhance generalizability. Studies exploring genetic and metabolic predictors of isotretinoin toxicity may enable more personalized treatment approaches. Additionally, evaluating long-term outcomes and recurrence patterns could provide further insights into optimizing treatment strategies. Future studies need to investigate the role that genetic and metabolic variation in retinoids induced adverse effects in case of isotretinoin. Furthermore, several models those used for regression analysis failed to reach statistical significance. However, several predictors were identified to be having clinical implications. Thus, these findings are preliminary and require cautious interpretation with validation from future perspective studies.

Conclusion

Isotretinoin remains a highly effective treatment for acne vulgaris in the Saudi population; however, its use is associated with a considerable burden of adverse effects requiring active clinical management. Multi-variable analysis identified cumulative dose (OR 3.75) and dose escalation during therapy (OR 93.71) strongest predictors of toxicity, while higher initial daily dose were protective against interventional requirements, while higher initial daily dose was protective against interventional requirements. The findings have particular relevance with regards to lipid abnormalities such as high LDL levels (9.9%) which may serve as useful, clinically relevant biomarkers for the degree of systemic in the study data. Although LDL levels were excluded from the independent variables used to determine clinical intervention within the regression model that determined when patients received clinical interventions, the future study will be required to establish if it should be given priority as one of the laboratory parameters monitored by clinicians along with others. Gender, body weight, and height associations are to be interpreted carefully, alongside comparing with previous research publications. This is because of a possible association with population-specific aspects that need further study using a larger prospective sample size.

Optimizing isotretinoin therapy requires a balance between efficacy and safety through individualized dosing, regular monitoring, and patient-centered care. Strengthening pharmacovigilance systems and enhancing patient education are essential to improving treatment outcomes and gradually increasing the dose could minimizing adverse effects in clinical practice. The findings present in the current study are consistent with the other published literatures, but require confirmation through prospective, multicenter study with higher sample size.

A possible enhancement to pharmacovigilance in Saudi Arabia could be the establishment of a national registry for isotretinoin, also the development of population-based educational materials for patients in Saudi Arabia. Follow up with all affected individuals is obligatory recommendation due to the high recurrence rate (19.4%) and to evaluate various strategies that can help maintain remission and decrease the possibility of relapse.

Funding Statement

This research was funded by the Deanship of Graduate Studies and Scientific Research at Jazan University, Saudi Arabia, under Project No. RP2026-072.

Data Sharing Statement

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request and subject to institutional ethical regulations.

Ethics Approval and Consent to Participate

Ethical approval was obtained from the Institutional Review Board of King Khalid University (ECM-2025-0061). The requirement for informed consent was waived because of the retrospective design and anonymized nature of the data. The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.

Disclosure

The authors declare no conflicts of interest.