Background

Psoriasis (PsO) is defined as a chronic, immune-mediated, gene-based disease with an inflammatory background that affects the skin, semi-mucosa, and joints. When joints and surrounding structures are involved, patients are classified as having psoriatic arthritis (PsA) [1,2,3,4]. The prevalence of PsO may range from 0.5 to 11.8% around the world [5,6,7,8,9], while the prevalence of PsA amongst patients with PsO varies from 5.9 to 48%, according to the patient characteristics and classification criteria used [1, 3, 4, 8]. PsO manifestations may vary; however, plaque PsO (or psoriasis vulgaris) is the most frequent skin phenotype, affecting approximately 90% of patients with PsO [9]. The disease may also affect the scalp, joints, creases, or nails, even in patients without skin lesions [10].

Fifty to 80 % of patients with PsO have concurrent nail lesions [11,12,13], which can lead to functional impairment, pain and discomfort, and decreased quality of life and general well-being [14, 15]. Despite its significant prevalence (around 50% of patients), nail manifestations are often neglected in daily clinical practice, probably due to a lack of recognition of its impact on patients or its relevance as an indicator of disease extension [15].

Psoriatic arthritis leads to impairments in a patient’s life, decreasing functional capacity and quality of life, which also increases the burden of disease to society. This burden highlights the need for early diagnosis and timely treatment for all comorbidities. In this sense, nail disease has been reported as a relevant risk factor for PsA [16] and may be employed as an early diagnostic parameter among patients with PsO.

Imaging techniques such as ultrasonography (US) have been increasingly used to diagnose and to monitor clinical features of PsO and PsA [17,18,19,20]. US findings usually include measures of thickness of the nail bed and the ventral and dorsal plates, as well as loss of definition, morphologic changes, and blood flow disturbances [21, 22]. Power Doppler (PD) and spectral Doppler (SD) are US techniques that are used to visualize nail inflammation. PD semiquantitatively shows nail inflammation through the detection of increased flow in blood vessels, whereas SD calculates the resistive index (RI) using systolic and diastolic peak flows of small vessels, which expresses the resistance to blood flow in the nail bed [22, 23]. Despite the relevance of US, discordant data are available on the usefulness of Doppler techniques for the evaluation of nail disease in PsO and PsA. Thus, the aims of this review are (i) to investigate the usefulness of nail US for the diagnosis of nail disease in patients with PsO and PsA; (ii) to gather data about parameters obtained through Doppler techniques (PD and SD) indicating inflammation of the nail bed, including but not limited to RI and vascularization of the nail unit; and (iii) to observe the differences between PsO, PsA, and healthy controls in RI and morphologic changes.

Main text

Methods

A systematic search was performed using MEDLINE via PubMed and LILACS (Latin American and Caribbean Health Sciences Literature) in order to identify studies addressing the use of US in nail assessment in terms of variables relevant in the context of PsO and PsA, to meet the previously mentioned goals. Two search strategies using a combination of controlled vocabulary (MeSH and DeCs keywords, for Pubmed and LILACS, respectively) and text words were adopted, as shown in Table 1. Searches were performed until March 20, 2018.

Table 1 Search strategy
Full size table

Conference proceedings of relevant scientific meetings in rheumatology (European League Against Rheumatism and American College of Rheumatology, as selected by the authors) were also screened. Only studies published during the past 10 years were considered eligible. Language selection was made manually by the reviewers.

After applying the predefined search strategies, the records were screened by two different reviewers using the following inclusion criteria: i) observational or non-therapy interventional studies; ii) patients with PsO and/or PsA; iii) studies assessing the use of US for nail assessment; and iii) papers reported in English, French, Portuguese, and Spanish only. Studies were deemed non-eligible if they consisted of any of the following exclusion criteria: i) clinical trials of any phase or study design or ii) case reports.

Initially, it was planned that in cases of discordance, a third reviewer would be the responsible for the final decision to include a selected article or not. No disagreements were identified in the review process; therefore, this strategy was unnecessary. Data extraction was performed by the reviewers, using a data collection tool specifically designed for this review. Variables abstracted from individual studies were: author, year, study design, sample size, baseline disease (if applicable), primary and secondary aims (if applicable), US assessments performed, nail parameters described, results. Assessment of bias was based on the Joanna Briggs Institute Critical Appraisal Instrument for Studies Reporting Prevalence Data [24, 25]. The risk assessment tool is descriptive and does not provide scores.

Results

A total of 48 records were initially identified. After application of the eligibility criteria, 13 were selected and included in this review. In addition, five abstracts were manually identified in the conference proceedings searched (Fig. 1), which provided the final number of 18 studies analyzed.

Fig. 1
figure 1

PRISMA flowchart

Full size image

The main characteristics of the 18 studies included in this review are summarized in Table 2.

Table 2 Studies included in the review
Full size table

Assessment of bias

The Joanna Briggs Institute Critical Appraisal Instrument for Studies Reporting Prevalence Data was applied to all of the included studies. In terms of sample frame and sampling, most studies used a clinic-based approach and described how the potential participants were recruited. Sample sizes varied from 10 to 238 patients, and all of them relied upon convenience, without clearly stating a sample size calculation rationale. Study subjects were appropriately described in all included studies, and valid methods to determine the presence of the pathological condition were used and extensively described. The statistical analysis plan was deemed appropriate for the 15 studies [26,27,28,29,30,31,32,33,34,35,36,37,38,39,40]. Overall, the risk of bias was assessed as moderate to high due to the small sample size of the studies.

Main findings of included studies

After full text analysis of the 18 records included in the review, a wide range of variables and methodologic approaches was identified. Tables 3, 4, and 5 summarize the key features that are relevant to this systematic literature review. Due to the methodologic variability of the included studies (as shown in Table 2), comparability or further data pooling was deemed not feasible. The descriptive data regarding grey-scale features, presence of Doppler vascularity, and RI measurement of nails’ vascularization are presented below.

Table 3 Studies comprising measurements on gray-scalea
Full size table
Table 4 Studies comprising quantitative results on PDa
Full size table
Table 5 Studies comprising quantitative results on spectral Dopplera
Full size table

Gray-scale features of the nail by ultrasonography

Twelve studies reviewed gray-scale findings [26, 28,29,30,31, 33,34,35,36,37,38,39]. Three of these studies were only pictorial essays, which were purely descriptive [41,42,43]. The terminology that was used by different investigators to describe gray-scale findings widely varied across studies, such as “loss of definition,” “hyperechoic definition,” “fusion,” or “hyperechoic focal involvement of the ventral plate,” all of which are likely to correspond to the loss of trilaminar appearance. The normal nail plate was usually described as “two hyperechoic white bands surrounding an anechoic well defined layer in between” and the lack of visibility of the latter anechoic layer may technically be named using one of these definitions. Although it was impossible to compare studies in the absence of a uniform definition, there were consistently more nail lesions as measured using US in patients with PsA (46–54%) [29, 35] and PsO (48.8–77.8%) [33, 35,36,37] compared with healthy controls (10%) [37]. In addition, patients with clinical nail disease were consistently found to have more lesions on US (57/101 [56.4%] vs 6/68 [8.8%]; p < 0.0001) [37] and had more frequent ventral nail plate deposits (median, 17.72 [Q1–Q3 = 10.14–27.83] vs 4.65 [Q1–Q3 = 0.05–16.23]; p = 0.0410) [31]. US results have a good agreement with clinical assessment for nail disease (kappa value = 0.79 for PsA patients and controls; p < 0.001) [33] and also a strong correlation (chi-square test, 10.769 for PsA and osteoarthritis patients; p = 0.001) [29]. However, it was not possible to confirm that US was more sensitive to detect nail disease versus clinical assessment. While there was higher number of nails with US features in the absence of clinical findings, there were also patients with positive clinical nail disease and no US features [36]. Nails with a false negative US test had mainly mild lesions, such as onycholysis or pitting with lower modified nail psoriasis severity index (mNAPSI, a psoriatic nail grading instrument used to assess severity of nail matrix and bed PsO by area of involvement in the nail unit) than those with true (i.e., marked) abnormalities on US (median mNAPSI, 10 [1–56] vs 17 [1–50]; p = 0.03), with a moderate absolute agreement between US and clinical assessment (76.3% with κ = 0.52, p < 0.0001) [37].

The studies also investigated nail thickness using US, reporting an increased thickness of nail plate, bed, and matrix in patients with PsO and/or PsA compared with controls [26, 28, 30, 31, 37, 39] (Table 3). Marina et al. was not able to demonstrate a difference in nail bed thickness between patients with PsO and controls [31]. Comparing 2 groups with PsO, one with nail disease and other with scalp PsO and/or inverse PsO, Acquitter et al. (2016) reported that the former group presented with statistically higher nail matrix thickness than patients in the latter group [30]. It was not possible to identify in the studies a comparison between PsO and PsA patients in terms of nail bed thickness with statistical significant differences.

Presence of vascularity within the nail unit by ultrasonography

Nail bed PD signals were variable in both patients with PsO and PsA across the studies, with a range varying from 20 to 96% [27, 29, 30, 35, 44]. A high frequency of vascularisation was also observed in healthy controls, ranging from 20 to 81.6% [27, 35]. Some studies demonstrated increased blood flow in patients with PsO [31, 39]. Comparing patients with PsO plus nails disease and patients with scalp PsO and/or inverse PsO, a higher frequency of PD signal in the nail bed was found in the first group compared with the second group [30] (Table 4). PD signals were usually scored semiquantitavely on a scale between 0 and 3. Interestingly, PsO was associated with all grades of PD signal severity [31]. On the contrary, Aydin et al. (2017) reported that a diagnosis of PsO was associated with a less frequent severe (grade 3) PD signal on the nail bed than in healthy controls (healthy controls vs PsO, 65.8% vs 34.9%; p = 0.002, 27].

Resistive index measurements

Three studies assessed RI measurements in patients with PsO or PsA compared with controls (Table 5) [31, 34, 40]. According to two of these studies, patients with PsO presented with statistically higher Nailfold Vessel RI (NVRI) measurements than healthy controls [31, 40]. Mendonça et al. (2014) assessed RI measurements in patients with PsA and reported that patients with PsA had lower RI measurements in both the nail bed in transverse and longitudinal planes than controls (PsA, mean of longitudinal plane measurement, 0.50 ± 0.13; mean of transverse plane measurement, 0.48 ± 0.09; controls, mean of longitudinal plane measurement, 0.86 ± 0.41; mean of transverse plane measurement, 0.70 ± 0.16). In addition, RI measurements in the nail bed in the longitudinal plane were correlated with RI measurements in the nail bed in the transverse plane (r = 0.333; p = 0.013) and with duration of medication use (r = 0.578; p = 0.002) and was negatively correlated with the presence of PD in the nail bed (r = − 0.213; p = 0.038). RI measurements in the nail bed in the transverse plane were not correlated with the presence of PD in the nail bed, while the measure of nail bed was correlated with the trilaminar appearance of nail (r = 0.472; p = 0.023, 34].

One study evaluated the sensitivity and specificity of RI measurements in patients with PsA [32]. In this study, patients with PsA presented statistically significant lower RI measurements than controls (p < 0.001), with high sensitivity and specificity for RI measurements in PSA patients (receiver operating characteristic curve = 0.858; p < 0.01). Patients with PsA and no symptoms of nail involvement also had lower RI measurements. Considering a 0.395 cut-off point for RI measurements, the results showed that RI measurements < 0.4 points were associated with 100 and 99% of sensitivity and specificity, respectively, for ungueal inflammatory activity [32].

Discussion

This systematic review was conducted to evaluate the current knowledge about the use of US for the diagnosis of nail disease in patients with PsO and PsA. However, the heterogeneous methodologic approaches did not allow us to perform a comparison of studies.

Although US is a method of diagnosis widely used in clinical practice for several diseases including PsA, real-world data shows that the use of this technique for the diagnosis of nail disease is still scarcely investigated in the literature, probably reflecting that the techniques is not routinely used in patients with PsO and PsA. Enthesitis/enthesopathies, joint synovitis and effusion, bone changes, tenosynovitis, and dactylitis are the main pathologies examined by US in patients with PsO and PsA [21]. The selected studies were mainly focused on the assessment of parameters that can differentiate healthy subjects with and without PsO and patients with PsO and PsA with and without nail disease [28, 29, 31,32,33,34,35,36,37,38,39,40,41,42,43].

A lower Doppler signal in the nail bed was found as marker of nail disease in patients with PsO and PsA compared with healthy controls. However the selected studies showed a wide variability for the presence of Doppler signal in the nail unit, mostly due to differences in the US equipment sensitivity or other variables such as Doppler settings, experience of the observer, or room temperature [29, 31, 34, 35, 39].

Some of the secondary outcomes of this review were related to resistance in the nail bed, such as to assess data regarding artifacts that could alter the RI measurement in the nail bed, the use of resistance in the nail bed to characterize inflammation, and differences in RI measurements in the nail bed among patients with PsO and PsA. Four of the included studies reported the RI measurements in the nail bed with conflicting results among patients with PsO and PsA, indicating the need for further evaluation in future studies to better determine how to apply the measure in clinical practice, including potential differences among specific subgroups.

Regarding artifacts that could alter the RI, two studies have shown significant differences when patients with PsO were compared with healthy controls and also when patients with PsO were stratified by the presence of nail disease [31, 40]. El-Ahmed and colleagues (2011) tested whether there were significant differences on NVRI measurements among groups of individuals based on sex, age, family history of PsO, and Psoriasis Area and Severity Index scores and no associations were found [40]. Also not all studies assessing these parameters reached statistical significance. Thus, this aspect of the disease still needs to be further investigated.

Morphologic changes, such as the thickness of nail beds, and nail plate, seem to be important parameters to analyze [28, 31, 34, 35, 37,38,39, 41,42,43]. In fact, patients with PsA and PsO presented significantly higher nail bed and nail plate thickness than controls [28, 39]; however, no study was able to predict more severe disease or the development of PsA based on this unique parameter [38].

Our review have limitations that need to be addressed, particularly the number of databases assessed and language limits, adopted due to logistic restrictions. Despite these limitations, we consider that the review was able to gather relevant and updated data about the current knowledge about the use of US to assess nail disease in PsO and PsA patients and also highlight areas for further investigation.

Conclusion

In conclusion, a significant variability across studies assessing nail disease using US in patients with PsO and PsA was observed. Samples were very diverse in terms of severity, disease duration and age. The measurement of thickness was the most frequently assessed parameter. Conflicting results exist on the presence of Doppler signals in the nail unit. Further studies are needed for the evaluation of the diagnostic value of this technique.