We performed an observational study of a retrospective cohort that was reviewed and approved by the Research Ethics Committee of Lugo and Santiago. Our research was performed following the principles of the Declaration of Helsinki.

The study population comprised patients referred from dermatology clinics to the Phototherapy Unit of Complejo Hospitalario Universitario Clínico de Santiago de Compostela, Santiago de Compostela, Spain to begin narrowband UV-B phototherapy between February 1, 2009 and March 31, 2015.

We included patients with skin diseases that were likely to benefit from phototherapy, namely, psoriasis, atopic dermatitis, vitiligo, plaque parapsoriasis, primary T-cell cutaneous lymphoma, acute and chronic pityriasis lichenoides, prurigo, pruritus sine materia, metabolic pruritus, urticaria pigmentosa, lichen ruber planus, disseminated granuloma annulare, polymorphous light eruption, and solar urticaria. Patients aged under 14 years, for whom MED was not calculated, were excluded.

The MED was calculated using the phototherapy cabinet (UV 7002 system [Waldmann] equipped with 21 F79/120W-TL01 tubes [311nm]). We irradiated 8 areas (1.5cm) on an unaffected area of the back with progressively increasing doses (increments of 20%) 1 day before initiation of treatment. In patients with phototypes I and II, the corresponding doses were 0.24, 0.29, 0.35, 0.42, 0.50, 0.60, 0.72, and 0.86J/cm2. In phototypes III and IV, the doses were increased to 1.00 and 1.20J/cm2. In order to obtain the erythemal irradiance of the TL01 lamps, we used a double-monochromator spectroradiometer (MACAM- SR 2271, MACAM) to calculate the absolute irradiance (4.73mw/cm2) and the irradiance with the potential to induce erythema (0.26mw/cm2), both of which values are measured annually. The resulting quotient is the correction factor used to transform these absolute irradiated doses (J/cm2) into erythemal doses (mJ/cm2): 13.19, 15.94, 19.24, 23.09, 27.49, 32.99, 39.58, 47.28, 54.98, and 65.97mJ/cm2, respectively. These biological effective doses for erythema correspond to the doses of the solar simulators; low MED can be determined using the TL01 lamps.

We recorded sociodemographic variables, clinical history, diagnosis, phototype, and MED value. We also recorded the drugs administered, doses (starting, final, and total [J/cm2]), the number of sessions, and adverse reactions (mild erythema [asymptomatic]; moderate erythema [erythema and transient pruritus<24h]; and severe erythema [erythema and pruritus>24h]; and pruritus). All data were retrieved from the specific phototherapy histories and from the electronic clinical history system of Galicia (IANUS). A database was designed specifically for this study.

The retrospective cohort was divided into 2 groups depending on the classification of MED as low or normal. The threshold for a low MED was defined by Pérez Ferriols et al.12 using the solar simulator as source.

Thus, in individuals with phototype I, MED values <0.12J/cm2 correspond to an erythemal dose<7mJ/cm2; in patients with phototype II, MED values <0.34J/cm2 correspond to an erythemal dose <19mJ/cm2; in patients with phototype III, MED values <0.49J/cm2 correspond to an erythemal dose<27mJ/cm2; and in patients with phototype IV, MED values <0.69J/cm2 correspond to an erythemal dose<38mJ/cm2.

From a clinical viewpoint, the prevalence of individuals with a low MED is approximately 5%. Assuming a significance of 5%, statistical power of 80%, and a 4% loss of data from clinical records due to underrecording, it was necessary to include 302 individuals in order to achieve a precision of 2.50%. The sample size calculation was performed using Ene3.0.17

Bivariate tests were performed to compare both groups. Categorical variables were compared using the chi-square test. Continuous variables were compared using the t test or Mann-Whitney test depending on normality, which was determined using the Kolmogorov-Smirnov test (with a Lilliefors correction). The risk of experiencing an adverse reaction was determined using the odds ratio (OR) and its 95% confidence interval (95% CI). Logistic regression was used to evaluate risk. Variables with a P value lower than 0.2 were entered into the multivariate regression model. All analyses were performed using IBM SPSS Statistics Version 19 (IBM Corp).18 Statistical significance was set at P<.05.

The study population comprised 302 patients, of whom 54% were men. The median (IQR) age was 49 (34-60) years. The most common diagnosis was psoriasis (148 patients [49%]), followed by atopic dermatitis (36 [11.9%]), cutaneous T-cell lymphoma (18 [6.0%]), photodermatosis (11 [3.6%]), prurigo (12 [4.0%]), and lichen planus (4.0%). Other conditions were diagnosed in 65 patients (21.5%).

In 156 cases (51.7%), the patient was not taking concomitant drugs, 74 (25.5%) were taking 1 drug, and 72 (23.8%) were taking 2 or more drugs. A total of 101 patients (69.2%) were taking photosensitizing drugs (Table 1).

The sample was divided into 2 groups depending on whether MED was normal or low.12 MED was normal in 240 patients (79.5%) and low in 62 (20.5%).

We found statistically significant differences between the groups according to sex (normal MED 47.9% vs low MED 77.4% in men, P<.001) and age (median age was 47 [33-58] years in patients with a normal MED and 56 [35.5-70] years in those with a low MED [P=.003]). Statistically significant differences were also found in the distribution of diagnoses between the groups (P<.001). However, no differences were found in the number of drugs consumed (P=.071), even when these where photosensitizing drugs (P=.806), or in the combination of phototherapy and acitretin (P=.196). Adverse reactions (erythema, pruritus, polymorphous light eruption, and solar urticaria) were recorded in 117 patients (38.7%). MED was low in 45.2% and normal in 37.1%. Erythema presented with a clinically relevant difference in patients with a low MED (35.5%) compared with those with a normal MED (25.4%). However, this clinical relevance was not observed in the case of pruritus, since 17.9% of patients with a normal MED reported the condition compared with 21.0% of patients with a low MED. Table 1 shows the results of this analysis in more detail. As for phototype, in patients with phototype IV, the percentage of patients with a low MED is higher than in patients with a normal MED (30.6% vs 14.6%; P=.013).

In order to construct the multivariate models and evaluate the effect of adverse reactions, the variables we selected for adjustment were those that, on the one hand, proved to be clinically relevant for MED and that, on the other, had a P value <.02 in the bivariate analysis. The variables that fulfilled these criteria were sex, age, diagnosis, and phototype. As shown in Table 2, Model 1 assesses the risk of adverse reactions in patients with a low MED compared with those whose MED was normal. Models 2 and 3 show an individual analysis of the most common adverse effects, namely, erythema and pruritus. Patients with a low MED did not have a greater risk than those who had a normal MED (Model 1). Similarly, refining the analysis by the type of adverse reaction did not reveal a significant risk. The risk of erythema expressed as the odds ratio (OR) in patients with a low MED was1.68 (95% CI, 0.91-3.29) (Model 2) and that of pruritus was 2.04 (95% CI, 0.99-4.22) (Model 3).

These analyses revealed an unexpected finding with respect to the diagnosis. The bivariate analysis revealed that psoriasis exerted a protective effect against low MED (OR,0.31 [95% CI, 0.17-0.58]). This significance persisted in the multivariate analysis adjusted for age, sex, and phototype (OR,0.31 [95% CI, 0.16-0.58]).

Comparison of the mean values of the dose for each group showed that the starting, final, and total doses were significantly different (P<.001, P=.004, and P=.001, respectively). However, the increase in dose and number of sessions was not significant (P=.304) (Table 3). Finally, we determined whether the distribution of MED differed according to the season of the year this value was determined; however, no significant differences were observed (Fig. 1).

Figure 1.

Variation in MED by season. MED indicates minimal erythema dose.

(0.12MB).

The authors declare that no tests were carried out in humans or animals for the purposes of this study.

The authors declare that they have followed their institutional protocols on publication of patient data.

The authors declare that no private patient data appear in this article.