Validation and usability of modified palmoplantar psoriasis area and severity index in patients with palmoplantar psoriasis: A prospective longitudinal cohort study

Plain Language Summary

Palmoplantar psoriasis is a difficult to treat, but common subtype of psoriasis affecting the palms and soles. It may have a great impact on quality of life based on the disease severity. Till date, there is no validated scoring instrument to score its severity. Though modified palmoplantar pustular psoriasis area and severity index is the frequently used scoring system used to score the severity of PPP, it is not validated. We validated this scoring system, which may be used in clinical practice and in clinical trials.

Introduction

Around 2–3% of the world population suffers from the chronic skin condition psoriasis, which is inflammatory and also hyperproliferative.¹ In India, the prevalence of psoriasis according to various studies has been proposed to be between 0.8 and 8%.^2–5 It is associated with a greater risk of cardiovascular comorbidities and renal complications.^6–8 Psoriasis affecting the palms and soles is known as PPP, and it may occur with psoriasis elsewhere or maybe the only site involved. PPP can decrease a patient’s quality of life in comparison to moderate-to-severe plaque psoriasis of the same magnitude.^9,10 Various studies have reported that palmoplantar involvement was seen in 2.8–40.9% of psoriatic patients.¹¹ In a study of 3065 patients with psoriasis, its prevalence was found to be 17.4%.¹² This study was performed at a tertiary care center in North India. The cardinal symptoms of PPP are іtchіng, burning and painful fissures. Though palms and soles contribute a little share to the total body surface area, because of their cosmetic and occupational appeal, it can be singularly debilitating. Evidence-based practice has necessitated the evaluation of a patient with psoriasis based upon standard and necessary assessment of ailment severity including variables, that are quintessential to patients. In this light, using validated severity scores is important for both clinical practice and research. An important lacuna in studies on the treatment of PPP is the lack of a disease-specific validated severity score or index. In several studies, the PPP area and severity index (PPPASI) and its various modifications have been used to assess the disease severity and response to treatment but none of them was uniform and validated.^11,13–17 Thus, we conducted this study to validate the modified palmoplantar psoriasis area and severity index (m-PPPASI).

Materials and methods

This was a prospective cohort study conducted in the Department of Dermatology, Venereology and Leprology, Post Graduate Institute of Medical Education and Research (Chandigarh, India) from July 2016 to December 2017, after obtaining the Institutional Ethical Committee clearance. A total of 73 patients with PPP satisfying inclusion and exclusion criteria were recruited after obtaining written informed consent.

Results

The distribution of cases and baseline demographic data are summarised in Table 2.

The m-PPPASI score was validated in 73 patients with PPP and was assessed at baseline, 2 weeks, and 6 weeks, but only 58 patients could be assessed at 12 weeks, and the remaining 15 patients were lost to follow-up.

Responsiveness to change and interpretability

Only the paired observations of the rater AN (m-PPPASI at 2 and 6 weeks) were used for analysing responsiveness to change, to prevent biases incurred from multiple observations on the same subjects. Pearson correlation of the m-PPASI score at 2 and 6 weeks was r = 0.92 (p < 0.0001) (n = 73). Similarly, for DLQI score, the Pearson’s correlation at 2 and 6 weeks was r = 0.7 (p < 0.0001) (n = 73). Since the correlations were >0.6 for both DLQI and m-PPPASI, they were considered to be suitable for MCID analysis.

To measure interpretability, the smallest mean change (MCID-1) and smallest percentage reduction (MCID-2) in the m-PPASI score which could be clinically useful was calculated using receiver operating characteristic curve analysis from the data of one of the raters (AN) [Figure 1 and Table 6].

Close

Figure 1:

Receiver operating characteristic curves for MCID-1 and MCID-2. Receiver operating characteristic curve analysis to calculate MCID-1 (a) Correlation is between point reduction of m-PPPASI at 0–6 weeks using point reduction of DLQI as an anchor at 0–6 weeks (area under the receiver operating characteristic curve = 0.95) (b) Correlation is between point reduction of m-PPPASI at 0–12 weeks using point reduction of DLQI as an anchor at 0–12 weeks (area under the receiver operating characteristic curve = 0.90) and MCID-2 (c) Correlation is between percentage reduction of m-PPPASI at 0–6 weeks using percentage reduction of DLQI at 0–6 weeks as an anchor (area under the receiver operating characteristic curve = 0.62) (d) Correlation is between percentage reduction of m-PPPASI at 0–12 weeks using percentage reduction of DLQI at 0–12 weeks as an anchor (area under the receiver operating characteristic curve = 0.68)

Export to PPT

Table 6: MCID-1 and MCID-2 by receiver operating characteristic curve analysis

MCID-1	Change in score	Sensitivity	Specificity	AUC
m-PPPASI 0–6 weeks (n = 52)	1.3	0.96	0.76	0.95 (0.91–1.00)
	1.4	0.94	0.81
	1.5	0.94	0.85
m-PPPASI 0–12 weeks (n = 43)	2.2	0.95	0.66	0.90 (0.81–0.98)
	2.4	0.93	0.73
	2.5	0.90	0.73
MCID-2
m-PPPASI 0–6 weeks (n = 52)	32.5	0.65	0.52	0.62 (0.44–0.80)
	33.1	0.63	0.52
	33.5	0.50	0.67
m-PPPASI 0–12 weeks (n = 43)	39.8	0.69	0.40	0.68 (0.52–0.83)
	40.5	0.65	0.53
	41.9	0.62	0.53

MCID: minimal clinically important differences, m-PPPASI: modified palmoplantar psorіasis area and severіty index, AUC: area under the receiver operating characteristic curve

The MCID-1 values obtained for m-PPPASI based on receiver operating characteristic curve analysis using DLQI as the anchor was 1.55 (area under the receiver operating characteristic curve, [AUC] = 0.959) at week 6 and 2.40 at week 12 (AUC = 0.902), we, therefore, recommend that the integer 2 be used to define MCID-1 for m-PPPASI [Figures 1a and 1b]. Similarly, 33.1% (AUC = 0.625) reduction at 6 weeks and 40.5% (AUC = 0.681) reduction at 12 weeks, respectively defined MCID-2 for m-PPPASI, and for practical purposes we recommend a 35% reduction be defined as MCID-2 for m-PPPASI [Figures 1c and 1d].

Categorisation

Using DLQI-based banding of the skin disease as a standard; we endeavoured to find the DLQI equivalent cutoff points for m-PPASI by using the receiver operating characteristic curve analysis. The resulting m-PPPASI equivalents came to be <6.3 for mild PPP, <8.3 for moderate PPP and >18.5 for very severe PPP. Using these values, we propose the practical scores of 0–5, 6–9, 10–19 and 20–72 to be used for categorising mild, moderate, severe and very severe PPP, respectively. [Figure 2 and Table 7].

Close

Figure 2:

Receiver operating characteristic curves for categorisation. Receiver operating characteristic curve analysis to calculate cut-off points for m-PPPASI using (a) Using DLQI < 6 as an anchor point (area under the receiver operating characteristic curve = 0.78) (b) DLQI < 11 as an anchor point (area under the receiver operating characteristic curve = 0.79) (c) Using DLQI > 20 as an anchor point (area under the receiver operating characteristic curve = 0.95)

Export to PPT

Table 7: DLQI equivalent m-PPPASI cut off points

	DLQI < 6		DLQI < 11		DLQI > 20
	Mild PPP		Moderate PPP		Very severe PPP
m-PPPASI	Sensitivity (%)	Specificity (%)	Sensitivity (%)	Specificity (%)	Sensitivity (%)	Specificity (%)
5.7	60	72
6.3	80	68
6.7	80	66
8.0			77	64
8.3			82	64
8.7			88	61
18.0					92	90
18.5					92	92
18.75					83	92
AUC	0.78 (0.62–0.94)		0.79 (0.69–0.89)		0.95 (0.9–1.0)

DLQI: dermatology life quality index, m-PPPASI, modified Palmoplantar Psorіasis Area and Severіty Index, AUC: area under the receiver operating characteristic curve

Discussion

Due to the absence of a validated scoring system in PPP, many alternate methods have been utilised to score the disease severity in various studies done in the past which include assessment through the use of a series of photographs, physician’s global assessment, patient’s global assessment, etc. The fact that none of these methods have been validated or standardised for PPP is a major drawback. As a result, it is becoming challenging to directly compare the studies. In this study, we tried to validate m- PPPASI, so that a uniform and standardised measure could be made available that could be used across studies, and a comparison between them would be possible.

The good internal consistency of the m-PPPASI indicates that the items are well related to one another. It simply states that different observers may employ m-PPPASI at various times and it can be relied upon to make treatment decisions. Inter-rater agreement was also good between the three raters. The test-retest reliability of the measurement results at two distinct but near points in time by all three raters shows that m-PPPASI demonstrates minimal fluctuation with the same observer when the disease remains static.

According to the literature, an item is considered to have passed the test of content validity if its content validity index for items is more than 0.78; nevertheless, several of the items failed this test.²¹ As in, the term “desquamation” ws deemed inappropriate for measuring the intended purpose of m-PPPASI, leading to the conclusion that the m-PPPASI was not appropriate for the study population. This was because experts believed that vesicular lesions, rather than psoriatic plaques, typically desquamate during healing. Instead, they suggested using the term “scaling,” which was more appropriate for plaques. “erythema”, “infiltration” (plaque PPP), “pustulation” (pustule PPP), and “desquamation” taken together were also not found to comprehensively reflect the measurement of PPP as it was felt that the other features to consider would be “fissuring” and “scaling” to objectively measure PPP. Due to inadequate sample size for the calculation, construct validity could not be tested, and criterion validity could not be performed due to the lack of a suitable “gold standard” to correlate the instrument with. DLQI being a quality of life instrument was deemed inappropriate to correlate with. Neverthless, we could establish a positive correlation between the DLQI and the m-PPPASI on several fronts, where it was considered appropriate. This was partly due to the absence of other standardized and validated instruments. An instrument aiming to clinically measure the disease would find little to no utility if it cannot reflect the change in status of the disease it purports to measure. Hence, in studying the responsiveness to change of the m-PPPASI, we correlated it with DLQI and found that m-PPPASI varied in parallel with the DLQI. This suggests that the m-PPPASI is at least as robust as the DLQI in monitoring responsiveness. We found that a decrease or increase in the value of m-PPPASI by at least 2 (defined by MCID-1) would suggest to the physician an alteration in the disease characteristic. Similarly, the smallest percentage reduction or increment that this value 2 would represent was calculated to be 35%, which means that there should be at least a 35% change in the disease for the physician to feel a treatment response (defined by MCID-2). For this purpose, we took a predefined smallest mean change in DLQI of 4 to be a measure of correlation with m-PPPASI. To increase the accuracy of MCID estimation, the receiver operating characteristic approach was used to determine MCID-1 and MCID-2, in which DLQI was used as an anchor. Since it has been argued that anchor-based methods to determine MCID are more clinically relevant than distribution-based studies. However, we recognised that responsiveness and MCID may differ depending on the population and the context; consequently, a single MCID may not be adequate or applicable to all study populations. Therefore, multiple clinically relevant anchors are considered for confirming responsiveness and determining MCID.^24,25 We were only able to use DLQI as an anchor in our study because we were constrained by the absence of other relevant anchors which could be used in PPP; consequently, we anticipate that the MCIDs we have established will facilitate comparison in future m-PPPASI-based studies. Nevertheless, in order to support our conclusions, further investigations using other anchors, such as Psoriasis Global Assessment and Lattice-system Psoriasis Global Assessment, ought to be carried out.

An attempt at categorising m-PPPASI in a fashion similar to the DLQI banding was undertaken in this study and we propose that m-PPPASI scores of 0–5, 6–9, 10–19 and 20–72 be taken for grading PPP as mild, moderate, severe, and very severe, respectively. This is analogous to and is derived while correlating m-PPPASI with the DLQI banding of 0–5, 6–10, 11–20 and 21–30 as none to mild, mild to moderate, severe, and very severe or disabling, respectively.

We recommend this categorisation in a clinically controlled environment as a rough guide and not an exact grading system in actual clinical practice on which treatment decisions are based, for which, a case-to-case variation, tailoring, and the physician’s own experience may be more appropriate in guiding treatment decisions.

In Figure 3, the scoring of m-PPPASI in one patient with palmar psoriasis is illustrated as done by the three raters (AN, TN and Sunil Dogra (SD)).

Close

Figure 3:

Palmar psoriasis. m-PPPASI as calculated by the 3 raters in this patient’s left palm was as follows: AN = 11.4, 11.4, 3 and 0.6, TN = 11.4, 11.4, 2.7 and 0.7 and Sunil Dogra (SD) = 11.4, 11.4, 3 and 0.6 at weeks 0 (a), 2 (b), 6 (c) and 12 (d) respectively

Export to PPT

Limitations

Small sample size and single-centre validation were the major limitations. Palmoplantar pustulosis cases were not included, however, it is increasingly becoming clear that PPP and palmoplantar pustulosis maybe two different disease entities²⁶ and it may not be appropriate to use a single scoring system to score both these entities. m-PPPASI doesn’t objectively measure all characteristics of PPP such as “fissuring” and “scaling” which could also be taken into consideration.

Conclusion

This is the first pilot study done for validation of m-PPPASI in the form of establishing reliability, good inter-rater agreement, face and content validity. The index was also found to be responsive to change, easily interpretable, and categorisable. Hence, we propose that the severity of the disease can be uniformly measured with m-PPPASI in both day-to-day clinical practice and clinical trials. This study will help in further research into PPP, which has traditionally been difficult to deal within psoriasis.