This corrigendum affects the Mediterranean Diet Score. Although distributions and correlations were not significantly altered from the data in the original manuscript, the revised manuscript comprises of minor amendments to Table 2 and Supplementary Information 6. In the revised manuscript, the associations between the Mediterranean diet score and psoriasis severity are not statistically significant at multinomial regression level resulting in modifications to ORs, 95% CIs, and P values in Table 3, and Supplementary Information 8. Reference to the Mediterranean Diet Score is removed from Table 4, Supplementary Information 9, and Supplementary Information 10. This correction arises from an incidental finding regarding the ‘meat and poultry’ component of the Mediterranean Diet Score, which was subject to a formula error.
Revised text and Table 2
For the MDS (score range 0-9) the mean (standard deviation) was 4.66 (1.72), for the DASH (score range 8-35) it was 23.88 (5.58), and for the PDI subtypes, the mean scores were 51.30 (7.29) for the oPDI, 52.29 (9.05) for the hPDI, and 51.64 (8.34) for the uPDI (score range 17-85 for all).
Only the uPDI reported no significant correlation with psoriasis severity. The remaining diet quality indices were negatively correlated with psoriasis severity: DASH (r −0.258, P < 0.001), hPDI (r −0.203, P = 0.001), MDS (r −0.192, P = 0.002) and oPDI (r −0.175, P = 0.005).
Table 2. The mean (standard deviation) of the DQIs across psoriasis severity categories and Pearson’s correlation coefficients with psoriasis severity
DQI = Diet Quality Index; MDS = Mediterranean Diet Score; DASH = Dietary Approaches to Stop Hypertension; oPDI = original Plant-based Diet Index; hPDI = healthy Plant-based Diet Index; uPDI = unhealthy Plant-based Diet Index.
The mean (standard deviation) of the DQIs are expressed as normalised values.
Psoriasis severity was determined using the self-assessed Simplified Psoriasis Index.
sa-SPI: ≤ 10 points (mild psoriasis); 10 – 20 points (moderate psoriasis); 20 - 70 points (severe psoriasis).
MDS: ≤ 3 points (low adherence); 3 – 6 points (modest adherence); 6 – 9 points (high adherence).
DASH: ≤ 8 points (very low adherence); 8 – 16 points (low adherence); 16 – 24 points (modest adherence); 24 – 32 points (high adherence); 32 – 40 points (very high adherence).
oPDI: ≤ 17 points (very low adherence); 17 – 34 points (low adherence); 34 – 51 points (modest adherence); 51 – 68 points (high adherence); 68 – 75 points (very high adherence).
hPDI: ≤ 17 points (very low adherence); 17 – 34 points (low adherence); 34 – 51 points (modest adherence); 51 – 68 points (high adherence); 68 – 75 points (very high adherence).
uPDI: ≤ 17 points (very low adherence); 17 – 34 points (low adherence); 34 – 51 points (modest adherence); 51 – 68 points (high adherence); 68 – 75 points (very high adherence).
Revised Supplementary Information 6
Supplementary Information 6. the mean (standard deviation) for sa-SPI and MDS tertiles, and DASH and PDI quintiles
DQI = Diet Quality Index; sa-SPI = self-assessed Simplified Psoriasis Index; MDS = Mediterranean Diet Score; DASH = Dietary Approaches to Stop Hypertension; oPDI = original Plant-based Diet Index; hPDI = healthy Plant-based Diet Index; uPDI = unhealthy Plant-based Diet Index. n 1 excluded from rank transformation for sa-SPI (59 points), n 1 excluded from rank transformation for oPDI (73 points), n 1 excluded from rank transformation for hPDI (76 points), and n 1 excluded from rank transformation for uPDI (73 points).
Revised text and Table 3
When adjusted for age, sex, smoking, AUDIT-C, energy intake, and psychological morbidity, very low adherence to the DASH (OR = 3.75, 95% CI 1.313 – 10.700, P = 0.01), and hPDI (OR = 4.04, 95% CI 1.251 – 13.064, P = 0.02) was associated with an increased likelihood of reporting higher psoriasis severity relative to very high adherence.
Table 3. Diet quality indices and the unadjusted and adjusted OR (95% CI) for psoriasis severity
Results of the multinomial regression were expressed as Odds Ratios (OR) with 95% Confidence Intervals (CI).
MDS = Mediterranean Diet Score; DASH = Dietary Approaches to Stop Hypertension; oPDI = original Plant-based Diet Index; hPDI = healthy Plant-based Diet Index; uPDI = unhealthy Plant-based Diet Index.
The reference categories for the diet quality indices were “very high adherence” (DASH and PDIs) and “high adherence” (MDS).
Confounder adjustments: Model VI = age (continuous), sex (male/female) and smoking (yes/no), Alcohol Use Disorders Identification Test Consumption score (continuous), energy kcal/day (continuous), and psychological morbidity (yes/no).
Model V = model VI and body mass index (continuous).
sa-SPI tertiles: T1 (low severity) ≤ 7; T2 (increasing severity) 8 - 17; T3 (high severity) ≥ 18.
MDS tertiles: T1 (low adherence) ≤ 3; T2 (modest adherence) 4 - 5; T3 (high adherence) ≥ 6.
DASH quintiles = Q1 (very low adherence) ≤ 16; Q2 (low adherence) 17 - 20; Q3 (modest adherence) 21 - 24; Q4 (high adherence) 25 - 27; Q5 (very high adherence) ≥ 28.
oPDI quintiles = Q1 (very low adherence) ≤ 43; Q2 (low adherence) 44 - 47; Q3 (modest adherence) 48 - 51; Q4 (high adherence) 52 - 55; Q5 (very high adherence) ≥ 56.
hPDI quintiles = Q1 (very low adherence) ≤ 41; Q2 (low adherence) 42 - 47; Q3 (modest adherence) 48 - 52; Q4 (high adherence) 53 - 57; Q5 (very high adherence) ≥ 58.
uPDI quintiles = Q1 (very low adherence) ≤ 41; Q2 (low adherence) 42 - 48; Q3 (modest adherence) 49 - 51; Q4 (high adherence) 52 - 56; Q5 (very high adherence) ≥ 57.
Revised Supplementary Information 8
Supplementary Information 8. diet quality indices and the adjusted OR (95% CI) for psoriasis severity (Model I-III)
Results of the multinomial regression were expressed as Odds Ratios (OR) with 95% Confidence Intervals (CI).
MDS = Mediterranean Diet Score; DASH = Dietary Approaches to Stop Hypertension; oPDI = original Plant-based Diet Index; hPDI = healthy Plant-based Diet Index; uPDI = unhealthy Plant-based Diet Index.
The reference categories for the diet quality indices were “very high adherence” (DASH and PDIs) and “high adherence” (MDS).
Confounder adjustments.
Model I = age (continuous), sex (male/female) and smoking (yes/no).
Model II = Model I and Alcohol Use Disorders Identification Test Consumption score (continuous).
Model III = Model II and energy kcal/day (continuous).
sa-SPI tertiles: T1 (low severity) ≤ 7; T2 (increasing severity) 8 - 17; T3 (high severity) ≥ 18.
MDS tertiles: T1 (low adherence) ≤ 3; T2 (modest adherence) 4 - 5; T3 (high adherence) ≥ 6.
DASH quintiles = Q1 (very low adherence) ≤ 16; Q2 (low adherence) 17 - 20; Q3 (modest adherence) 21 - 24; Q4 (high adherence) 25 - 27; Q5 (very high adherence) ≥ 28.
oPDI quintiles = Q1 (very low adherence) ≤ 43; Q2 (low adherence) 44 - 47; Q3 (modest adherence) 48 - 51; Q4 (high adherence) 52 - 55; Q5 (very high adherence) ≥ 56.
hPDI quintiles = Q1 (very low adherence) ≤ 41; Q2 (low adherence) 42 - 47; Q3 (modest adherence) 48 - 52; Q4 (high adherence) 53 - 57; Q5 (very high adherence) ≥ 58.
uPDI quintiles = Q1 (very low adherence) ≤ 41; Q2 (low adherence) 42 - 48; Q3 (modest adherence) 49 - 51; Q4 (high adherence) 52 - 56; Q5 (very high adherence) ≥ 57.
Revised text and Table 4
The red and processed meat component of the DASH score was associated with psoriasis severity (R 2 0.059, β = 0.209, t = 3.328, P = 0.001), with greater intakes predicting more severe psoriasis and retained significance across all covariate adjustment models in the univariate linear regression (β = 0.190, P = 0.004), even after adjustment for BMI. On the other hand, the nuts and legume component of the DASH score was negatively associated with psoriasis severity (R 2 0.081, β = -0.153, t = −2.423, P = 0.02) with greater intakes predicting milder psoriasis and retained significance until adjustment for BMI where the association was no longer significant (β = −0.128, P = 0.06).
Table 4. Extracted DASH components as standardised predictors of psoriasis severity, followed by the results of the univariate regression analyses adjusted for covariate models I-V
DASH = Dietary Approaches to Stop Hypertension.
Stepwise multiple linear regression values are expressed as standardised β-coefficients, P values, R2 values and t values.
Univariate linear regression models adjusted for age (continuous), sex (male/female) and smoking (yes/no)(model I), model I and Alcohol Use Disorders Identification Test Consumption score (continuous)(model II), model II and energy kcal/day (continuous) (model III), model III and psychological morbidity (yes/no) (model IV), and model IV and body mass index (continuous) (model V).
Revised text and Supplementary Information 9
The mediation analysis (Supplementary Information 9) showed that BMI fully mediated the association between the hPDI, oPDI, and uPDI, and psoriasis severity, but partially mediated the inverse association with the DASH indicating an independent association between the DASH diet and psoriasis severity that is not dependent on BMI.
Supplementary Information 9. mediation analysis examining the mediating effect of BMI on the DQI and psoriasis severity associations
BMI = Body Mass Index; DQI = Diet Quality Index; sa-SPI = self-assessed Simplified Psoriasis Index; DASH = Dietary Approaches to Stop Hypertension; oPDI = original Plant-based Diet Index; hPDI = healthy Plant-based Diet Index; uPDI = unhealthy Plant-based Diet Index.
A 3-step mediation analysis was conducted. Step 1 = linear regression with DQI (as the independent variable) and BMI (as the dependent variable) for the exposure-mediator effect. Step 2 = a multiple linear regression with DQI and BMI (as the independent variables) and sa-SPI (as the dependent variable) for the direct effect. Step 3 = linear regression with DQI (as the independent variable) and sa-SPI (as the dependent variable) for the total effect. Partial mediation = if all unstandardised - β coefficients for a given DQI are statistically significant and the unstandardised - β coefficient of the DQI in the direct effect model is closer to zero than that of the DQI in the total effect model. Full mediation = if the unstandardised - β coefficient for a given DQI is not statistically significant within the direct effect model, but the mediator is statistically significant. The effect of the oPDI, and hPDI on sa-SPI is fully mediated by BMI and disappears when the BMI is added to the regression model. BMI partially mediates the effect of the DASH on sa-SPI meaning that when BMI is added to the model, the effect of DASH is reduced but not nullified.
Revised Supplementary Information 10
Supplementary Information 10. extracted DASH components as unstandardised predictors of psoriasis severity, followed by the results of the univariate regression analyses adjusted for covariate models I-V
DASH = Dietary Approaches to Stop Hypertension.
Stepwise multiple linear regression values are expressed as unstandardised β-coefficients (95% confidence intervals), P values, R2 values and t values.
Univariate linear regression models adjusted for age (continuous), sex (male/female) and smoking (yes/no)(model I), model I and Alcohol Use Disorders Identification Test Consumption score (continuous)(model II), model II and energy kcal/day (continuous) (model III), model III and psychological morbidity (yes/no) (model IV), and model IV and body mass index (continuous) (model V).
Revised text in discussion
Participants with a lower adherence to healthy dietary patterns such as the DASH, hPDI, and oPDI, were at least twice as likely to report the highest psoriasis severity.
Adherence to the MD was inversely correlated with psoriasis severity but was not a predictor of disease severity, conflicting with the results of the NutriNet Santé cohort study [10]. The NutriNet Santé study involved a larger prospective cohort than the APPLE study, and was conducted in a French population where MD adherence was more likely and lacks generalisability to northern European countries such as the UK where the MD is not the traditional eating pattern. The APPLE study however, classified psoriasis severity using a validated tool, the sa-SPI, whereas the NutriNet Santé study used a combination of self-rated severity, hospitalisation history, and medication use as a proxy for severity levels, which may be less accurate.
We identified the nuts and legume component of the DASH score that was associated with likelihood of reporting milder psoriasis. This could be linked to: (i) the anti-inflammatory properties of a range of (poly)phenols, micronutrients, and fatty acids [35-37], and (ii) the insoluble fibre contents of these foods, which may exert immunomodulatory activity through the synthesis of short chain fatty acids upon fermentation by the host microbiota [38]. In an Italian cohort, olive oil and fish emerged as protective foods for severe psoriasis as per the PREDIMED questionnaire [11].
Open access
