Differences in Clinical Responses to Ustekinumab Treatment among Body Regions: Results from a Real-World Prospective, Observational, and Multi-Center Study in Korea.

BACKGROUND: In psoriasis treatment, not all body regions improve simultaneously after clinical interventions.
OBJECTIVE: This study was aimed at evaluating clinical responses across body regions, which may differentially influence patient treatment plans.
METHODS: This prospective, observational, and multi-center study was conducted in Koreans who adhered to ustekinumab treatment based on criteria per local label and reimbursement guidelines. A total of 581 were included in this analysis.
RESULTS: The mean (±standard deviation) psoriasis area severity index (PASI) score at baseline, age, disease duration, and body surface area (%) were 18.9±9.69, 44.2±13.29 years, 11.3±9.65 years, and 27.8±17.83, respectively. Across the head and neck, upper extremities, trunk, and lower extremities, the correlation between the PASI sub-scores for the upper and lower extremities was the highest (r=0.680). The mean PASI sub-score for the lower extremities was the highest at baseline. PASI90 and PASI100 scores were the highest for the head and neck region, indicating the highest response rates, while those for the lower extremities were consistently low at all visits.
CONCLUSION: We found differences in regional ustekinumab responses, with the lower extremities being the most difficult to treat. These findings should be considered in psoriasis treatment.

INTRODUCTION

Psoriasis is a chronic inflammatory disease characterized by plaques that display erythema (redness), infiltration (thickness), and desquamation (scaliness) on the skin due to a systemic immune system abnormality and may affect many aspects of a patient’ life1. Ustekinumab (Stelara®; Janssen-Cilag International NV, Beerse, Belgium) is a fully human monoclonal antibody that inhibits interleukin (IL)-12 and IL-23 by binding to their shared p40 protein subunit23. Ustekinumab represents one of a number of biologics that have been approved and demonstrated improved treatment outcomes for psoriasis over the past 10 years4.

The psoriasis area severity index (PASI) is the standard measurement most often used to determine treatment efficacy in many clinical trials. The PASI evaluates four body regions (the head/neck, upper limbs, trunk, and lower limbs) and three plaque characteristics (erythema, induration/thickness, and scaling). The total PASI score is calculated by summing the regional scores weighted by surface area of the skin5. PASI scores are typically reported as a composite measurement that does not account for results for each individual body region. The purpose of this study was to assess and compare the characteristics of and clinical responses to ustekinumab across the four PASI body regions independent of weighting by surface area of each body region. Further, clinical responses across body regions may differentially affect patients’ quality of life (QoL) and impact communication with patients for formulating their treatment plans.

MATERIALS AND METHODS

Study design

This was a post-hoc analysis of a postmarketing surveillance study for ustekinumab conducted in Korea6. The study enrolled patients aged 18 years or higher to whom ustekinumab was administered for the treatment of psoriasis. Demographic variables (age, sex, body mass index [BMI], and smoking/drinking history), disease-related variables (disease duration, age at diagnosis, comorbidity, and medication history), and disease severity variables (body surface area [BSA], PASI, and physician’s global assessment [PGA]) were collected. Data on PASI, as a measure of effectiveness, were collected at baseline and each patient visit. The Institutional Review Board of each participating site reviewed and approved the study protocol (IRB number of the corresponding author’s hospital [Seoul National University Bundang Hospital]: B-1112/141-201). Written informed consent was obtained from all participating patients.

Statistical analysis

Descriptive statistics for continuous variables were presented as means with standard deviation (SD), and dichotomous variables were presented as frequencies with percentages in parentheses. Only patients who adhered to treatment based on criteria per local label and reimbursement guidelines were included in the effectiveness assessment. For example, local practice guidelines for ustekinumab stipulate that patients should have a PASI ≥10 and BSA ≥10% at baseline and that treatment after the initial injection should be according to the following dosing intervals; visit 2: 1 month (or 4 weeks)±2 weeks after baseline, and visits 3~6: 3 months (or 12 weeks)±2 weeks after the previous visit. Scores for each of the four PASI body regions were extracted for comparison between the body regions. Pearson’s correlation coefficient was used to determine the linear relationship between the baseline PASI sub-score for each body region.

To compare effectiveness across the four body regions, a linear mixed-effects model was used with subject-specific intercepts as random effects, an unstructured covariance structure for continuous outcomes, generalized estimating equation with logit link, and compound symmetry correlation for binary outcomes (recommended for analyzing longitudinal or correlated data). The confounding covariates of age and sex were included as fixed effects in all models.

For comparisons of effectiveness between the four body regions at each post-baseline visit, the assessment of PASI response (PASI75, 90, and 100) included only patients with PASI sub-scores of ≥1 for the body regions at baseline. Furthermore, determination of the proportion of patients achieving a score of 0 for each of the clinical signs (erythema [redness], infiltration [thickness], and desquamation [scaliness]) assessed in the PASI included only patients with a baseline score of ≥2 (moderate to very marked) for each respective body region. Severity at baseline (reflected by body region PASI or clinical signs scores) was used as a covariate. All statistical tests were performed using two-sided tests, and p-values <0.05 were considered statistically significant. All analyses were performed using the statistical software package SAS 9.4 (Statistical Analysis System; SAS Institute, Cary, NC, USA).

RESULTS

Baseline characteristics

In all, 581 patients with psoriasis who adhered to ustekinumab treatment based on criteria per local label and reimbursement guidelines were included in the analysis. The mean (±SD) age of the patients was 44.2±13.29 years, and 391 patients (67.3%) were male. The mean BMI, disease duration, and age at diagnosis were 24.3±3.58 kg/m2, 11.3±9.65 years, and 32.9±15.07 years, respectively. Regarding psoriasis-associated comorbidities, 60 (10.3%), 37 (6.4%), 31 (5.3%), 27 (4.6%), and 20 (3.4%) patients had hypertension, type 2 diabetes mellitus, obesity, psoriatic arthropathy, and dyslipidemia, respectively. Furthermore, 60.6% (n=352) of the patients did not have a smoking history, and 39.1% (n=227) had no history of drinking at baseline. The mean BSA (%) and PASI were 27.8±17.83 and 18.9±9.69, respectively. Regarding prior treatment, 83.8% (n=487) of the patients had been treated with phototherapy, 84.7% (n=492) had received oral agents, and 9.6% (n=56) had previously used biologics (Table 1).

Table 1

Baseline characteristics*

Category		No. of available patients	Value†
Age (yr)		581	44.2±13.29
Sex		581
	Male		391 (67.3)
	Female		190 (32.7)
Body mass index (kg/m2)		578	24.3±3.58
Disease duration (yr)		566	11.3±9.65
Age at diagnosis (yr)		566	32.9±15.07
Comorbidity‡		581
	Hypertension		60 (10.3)
	Type 2 diabetes mellitus		37 (6.4)
	Obesity		31 (5.3)
	Psoriatic arthropathy		27 (4.6)
	Dyslipidemia		20 (3.4)
	Hepatic steatosis		10 (1.7)
	Tinea pedis		9 (1.5)
	Liver function test abnormal		7 (1.2)
	Osteoporosis		6 (1.0)
Smoking history		581
	Never		352 (60.6)
	Current		132 (22.7)
	Former		97 (16.7)
Drinking history		581
	Never		227 (39.1)
	Current		257 (44.2)
	Former		97 (16.7)
Disease-specific variables
	BSA (%)	581	27.8±17.83
	PASI	581	18.9±9.69
PGA≥‘moderate’		581	508 (87.4)
Prior treatment		581
	Phototherapy		487 (83.8)
	Oral agent		492 (84.7)
	Biologics		56 (9.6)

Values are presented as mean±standard deviation or number (%). BSA: body surface area, PASI: psoriasis area severity index, PGA: physician’s global assessment. *Total number of patients in the analysis set was 581. †Percentages are based on the number of available patients per variable. ‡Comorbidity includes lists with a frequency of 1% or higher.

Correlations between the four PASI body regions (the head and neck, trunk, upper extremities, and lower extremities) at baseline are presented in Table 2. The head and neck showed fair or moderate correlation, in the range of 0.323~0.447, with the other body regions. However, the correlation between the trunk and upper extremities or the trunk and lower extremities showed a positive linear relationship (r=0.638 and 0.606, respectively), and the correlation between the upper and lower extremities was the highest (r=0.680). Comparison of the mean baseline PASI sub-scores revealed significant differences between the body regions, with the highest being for the lower extremities. Furthermore, the mean PASI sub-score for the lower extremities was also the highest in each of the five stratified total PASI score groups (Table 3). In addition, there were significant differences in the proportions of patients with involvement of each clinical sign between body regions, with the highest proportions showing involvement of the lower extremities for each sign (Table 3).

Table 2

Correlation between body regions (n=581)

Pearson’s correlation coefficient (r)	PASI (H)	PASI (T)	PASI (U)	PASI (L)
PASI (H)	1.000
PASI (T)	0.406	1.000
PASI (U)	0.447	0.638	1.000
PASI (L)	0.323	0.606	0.680	1.000

PASI: psoriasis area severity index, H: head and neck, T: trunk, U: upper extremities, L: lower extremities. All p-values are <0.0001.

Table 3

PASI score and clinical sign by body regions at baseline

Category			No. of available patients	Body region*				p-value
				Head and neck	Trunk	Upper extremities	Lower extremities
PASI
	Body region score at baseline		581	12.9±11.38	19.0±11.78	17.5±11.16	21.1±11.70	<0.0001
	Body region score at baseline by the total PASI group
		10≤total PASI <15	269	9.0±8.22	12.3±5.32	11.1±5.11	13.5±4.62	<0.0001
		15≤total PASI <20	134	12.2±8.78	16.6±5.79	15.8±6.22	19.3±4.67	<0.0001
		20≤total PASI <25	61	12.8±9.38	21.8±7.89	19.7±6.13	26.1±7.46	<0.0001
		25≤total PASI <30	48	17.5±11.53	28.5±8.02	25.6±7.19	29.1±6.35	<0.0001
		30≤total PASI	69	26.3±16.16	41.2±12.00	38.4±12.41	44.2±10.82	<0.0001
Points for clinical sign ≥2 points (moderate)
	Erythema (redness)		581	395 (68.0)	535 (92.1)	521 (89.7)	556 (95.7)	<0.0001
	Infiltration (thickness)		581	298 (51.3)	450 (77.5)	461 (79.3)	487 (83.8)	<0.0001
	Desquamation (scaliness)		581	288 (49.6)	425 (73.1)	423 (72.8)	477 (82.1)	<0.0001
Point of area ≥2 points (10%~29%)			581	400 (68.8)	520 (89.5)	505 (86.9)	558 (96.0)	<0.0001

Values are presented as mean±standard deviation or number (%). PASI: psoriasis area severity index. PASI score’s p-values pertain to the ‘body region’ effect from the linear mixed-effects model with ‘age’, ‘sex’, and ‘body region’ as fixed effects; ‘intercept (i.e., subject-specific)’ as a random effect; and unstructured covariance as covariance structure. Clinical sign’s p-values pertain to the ‘body region’ effect from the generalized estimating equation with ‘age’, ‘sex’, and ‘body region’ as covariates; ‘body region’ as a repeated measure; and compound symmetry correlation as working correlation. *Percentages are based on the number of available patients per variable.

Effectiveness over time between the body regions

With regard to the PASI75 response, the head and neck scores generally indicated the lowest response rates through visit 6 among the four body regions (Fig. 1A). In addition, significant differences in PASI90 and PASI100 responses were noted between the body regions at all visits. Ironically, the head and neck region showed the highest PASI90 and PASI100 response rates, while those for the lower extremities were markedly lower at all visits (Fig. 1B, C). Comparison of improvements in clinical signs indicated significant differences in erythema and infiltration between the body regions at all visits. Additionally, the lower extremities showed the lowest rates of achieving a score of 0 at all visits for all clinical signs (Fig. 2), and erythema (redness) showed the lowest response rates across clinical signs in all body regions at all visits (Fig. 3)

Fig. 1

Psoriasis area severity index (PASI) response by body region. (A) PASI75 response; (B) PASI90 response; and (C) PASI100 response. Overall p-value of ‘body region’ *p<0.05, **p<0.01, ****p<0.0001.

Fig. 2

Clinical sign response by body region; (A) erythema (redness), (B) infiltration (thickness), and (C) desquamation (scaliness). Overall p-value of ‘body region’ *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Fig. 3

Response in body regions by clinical sign. No signs (0 point) in head and neck (A), trunk (B), upper extremities (C), and lower extremities (D). Overall p-value of ‘body region’ *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

DISCUSSION

Patients who achieve a PASI75 or even a PASI90 response using biologics show persistent residual psoriatic lesions in certain areas. If patterns in areas in which lesions persist despite biological treatment can be identified, patient expectations may be set more accurately during the course of care. Overall, 581 patients who adhered to local label and reimbursement guidelines for ustekinumab treatment from among 977 patients with psoriasis who were enrolled in this postmarketing surveillance study in Korea were included in the analysis to address regional differences in the extent of disease at baseline and response to ustekinumab treatment.

With regard to the degree and severity of psoriasis across the four body regions at baseline, correlation between the head and neck and the other body regions was low (r=0.323~0.447). In particular, correlation between the head and neck and the lower extremities was the lowest (r=0.323). Comparison of PASI sub-scores across the four body regions at baseline showed that those for the head and neck were the lowest (12.9±11.38) and those for the lower extremities were the highest (21.1±11.70). Similarly, comparison of the severity of clinical signs across the four body regions at baseline showed that the proportions of patients with moderate to very marked (score of ≥2) erythema (redness), infiltration (thickness), and desquamation (scaliness) were the lowest for the head and neck region and highest for the lower extremities (68.0% vs 95.7% for erythema; 51.3% vs 83.8% for infiltration; and 49.6% vs 82.1% in desquamation, respectively). Therefore, differences in disease burden by region should be considered when evaluating the overall degree of severity of psoriasis in individual patients.

The disease burden of facial psoriasis is significant7; scalp psoriasis has also been reported to have a negative impact on patients’ QoL and affect patient health physically and psychologically regardless of psoriasis severity89. Previous reports indicate that the body region with the greatest negative impact of psoriasis on QoL and highest disease burden is the head and neck. However, our results show that the most severely affected body region based on objective PASI assessments is the lower extremities rather than the head and neck, which implies that QoL and disease severity are not necessarily related and their relationship may be affected by the body regions involved. Plaque-type psoriasis is known to develop commonly at the elbows, knees, and scalp; however, it can affect other parts of the body as well10. Larko11 reported that the scalp, nail, sole, and intertriginous area were particularly difficult areas to treat12. Patients with psoriatic lesions at such sites may experience a psychological burden because of their visibility13 and the negative impact on work productivity14.

Our results showed that PASI90 and PASI100 responses were the highest for the head and neck, the body region with the greatest impact of psoriasis on patients’ QoL. To our knowledge, these findings are the first to suggest that the head and neck region may tend to show greater improvements in psoriasis compared to other body regions, although head and neck psoriasis may be resistant to biologic treatment in some patients. This may be reflected by our finding that the head and neck region generally showed numerically lower PASI75 response rates relative to those associated with other body regions. This means, in other words, although head and neck region is difficult part to achieve PASI75 response among 4 body regions, however, most of the patients are expected to show high responses (PASI90 & 100 response) if the patients meet clinically meaningful response cut-off (PASI75 response). In addition, with regard to the three clinical signs assessed in the PASI, the lower extremities consistently showed the least improvement at each visit through visit 6 (53.7±2.1, weeks), which also supports that the lower extremities were the most difficult body region to treat. This is consistent with recent findings that the most common sites of recalcitrant psoriasis are on the lower legs15. Therefore, in addition to treating psoriatic lesions on exposed body regions including the head and neck, which affect patients’ QoL, attention should be paid to the treatment of the most difficult to treat body region, the lower extremities. Besides, as one of the limitations as well as further study concepts, other specific sites beyond comprehensive 4 body regions should be further investigated. For example, nails, scalp, palms, and soles are known as a part difficult to treat16. However, we didn’t specify those parts in our study and couldn’t see treatment outcomes in those specific sites.