Obsessive-compulsive symptoms in first episode psychosis and risk states: Systematic review with meta-analysis

Background and hypothesis: Recent studies have reported high prevalences of obsessive-compulsive symptoms and obsessive-compulsive disorder in at risk and first-episode psychosis patients. This sparked an interest in the effect of these symptoms in the clinical characteristics and outcomes of patients. However these studies have never been formally meta-analyzed. Study design: Systematic review and meta-analysis of prevalence of obsessive-compulsive symptoms and obsessive-compulsive disorder in at risk and first-episode psychosis patients and comparison of clinical characteristics and outcomes in patients with and without obsessive-compulsive symptoms. Study results: Obsessive-compulsive disorder was present in 7.9 % (5.9 to 10.0 %) and 10.5 % (8.3 to 12.8 %) and obsessive-compulsive symptoms in 21.4 % (8.3 to 38.2 %) and 34.0 % (26.3 to 42.1 %) of at risk and first episode psychosis patients respectively. The prevalences of obsessive-compulsive symptoms had high heterogeneity due in part to different measurement methods and cut-off values. Similar ages of onset for OCS and psychosis symptoms were found (mean difference (cid:0) 0.49 years, 95 % CI -1.74 to 0.77). Patients with obsessive-compulsive symptoms had statistically insignificant higher Positive and Negative Syndrome Scale (positive subscale) scores and marginally higher depression scores. There were no differences between both groups in age of onset, Positive and Negative Syndrome Scale (negative subscale) score, risk of conversion to psychosis, anxiety score, suicide rate, and functionality score. Conclusions: Obsessive-compulsive disorder and obsessive-compulsive symptoms are very prevalent in at risk and first-episode psychosis patients.


Introduction
The co-occurrence of obsessive-compulsive symptoms (OCS) and psychosis has been known for at least a century (Gordon, 1926). The occurrence of OCS and obsessive-compulsive disorder (OCD) in patients with schizophrenia is well described, with prevalence rates of 12.3 % for OCD and 30.7 % for OCS in Swets et al. (2014), however with high heterogeneity. Swets et al. (2014) investigated the motives for high heterogeneity in the prevalence of OCS, and determined that the use of different cut-off points in OCS assessment scales had a significant role in it. This is much larger than that found in the general population (3.5 % for OCD and 8.7 % for OCS) (Angst et al., 2004). Earlier authors argued that psychosis in patients who also had OCS were less severe, (Stengel, 1945). Later research pointed somewhere else, with evidence that OCS in schizophrenia were associated with greater severity of positive and negative symptoms (Cunill et al., 2009), greater need of inpatient admission, worse functionality (Berman et al., 1995;Hwang et al., 2000) and greater risk of suicide (Sevincok et al., 2007).
The relationship between OCS and schizophrenia is unclear. Changes to the DSM-IV allowed for the co-diagnosis of both disorders, and some authors have worked on schizo-obsessive disorder as a subtype of schizophrenia (Hwang et al., 2009;Poyurovsky et al., 2012). On the other side, both disorders may co-occur independently. Most studies on OCS and OCD in psychosis have been conducted in patients with chronic schizophrenia. That presents some limitations, such as the retrospective evaluation of disease onset and the contributing effect of antipsychotic agents on OCS (Schirmbeck et al., 2011).
In an attempt to prevent and better treat psychotic disorders, the focus of diagnosis and management has shifted towards earlier stages of psychosis (Yung et al., 2004). Attention has been paid to two groups of patients that were here included in the "early psychosis" designation. 1) Those with first episode psychosis (FEP), in which formal criteria for schizophrenia or a schizophrenia spectrum disorder have been met. And 2) those who do not meet formal criteria, but present subsyndromal symptoms of some kind, that will be called "at risk" in the remainder of this article. Different criteria were developed to operationalize at risk states. The ultra high risk (UHR) criteria, from the PACE clinic in Australia (Yung and Nelson, 2013), include three groups of patients: the attenuated psychotic syndrome group (those who hava had subthreshold positive symptoms during the past year); the brief limited intermittent psychotic symptom group (those who have had episodes of frank psychotic symptoms for less than a week and spontaneously remitted); and the trait and state risk factor group (those with a first-degree relative with a psychotic disorder or who have schizotypal personality disorder and a decrease in functioning or chronic low functioning during the previous year). These criteria were reoperationalized by the PRIME clinic in the USA, that developed the concept of psychosis risk syndrome (Miller et al., 2003). Finally, the German Research Network on Schizophrenia developed the concepts of early initial prodromal state (selfexperienced cognitive thought and perception deficits, also called basic symptoms according to Huber and Gross, 1989, or a clinical decline in functioning in combination with well established risk factors) and late initial prodromal state (similar to the UHR criteria) (Bechdolf et al., 2005).
There were early 20th century descriptions of OCS in the beginning stages of psychosis, including by Kraepelin (Kraepelin, 1919). Some recent studies have reported high prevalences of OCS and OCD in FEP and at risk patients. This sparked an interest in the effect of OCS in the clinical characteristics and outcomes of patients in early stages of psychosis, such as severity of psychosis, severity of comorbidities, temporal association between OCS and psychotic symptoms and, most notably, risk of conversion to psychosis, with varying results. These studies may shed much-needed light on the relationship between psychosis and OCS, as they bypass some limitations of chronic schizophrenia studies. However, they have never been formally summarized.
Our objectives were to conduct a systematic review of the literature of OCD and OCS in FEP and at risk patients and a meta-analysis of 1) their prevalence and 2) time of onset of OCS in relation to psychotic symptoms and 3) clinical characteristics of patients who present with OCS (OCS+) in relation to those who don't (OCS-).

Material and methods
This systematic review and meta-analysis was conducted according to the PRISMA guidelines (Page et al., 2021). A registration was submitted to PROSPERO with the registration number 344066 and was rejected because PROSPERO currently focuses on COVID-19 submissions. That submission was titled Obsessive-Compulsive Symptoms in First-Episode Psychosis: Systematic Review and Meta-Analysis. We later decided to also include at risk patients since those two states (FEP and at risk) are deeply related.
MEDLINE, CENTRAL and Psycnet were searched from inception to February 2023. Reference lists were cross-checked for additional references. We used the following search terms: ( ). Titles, abstracts and full texts were screened independently by 2 reviewers. Disagreements were solved by consensus. Original studies of any methodology that included newly published data on the prevalence of OCS or OCD in patients with early psychosis or on clinical characteristics were included. Studies were accepted regardless of patient age or comorbidity and regardless of publication year. We excluded publications on other illnesses, on psychosis with no information on OCS, on OCS in early psychosis but no information on prevalence or clinical characteristics, or on other psychiatric symptoms. When more than one study reported on the same cohort, we only accepted the most complete study, or the most recent. The primary outcome was the prevalence of OCD in patients with early psychosis. Secondary outcomes were the prevalence of OCS in patients with early psychosis; age of onset of OCS in relation to age of onset of psychotic symptoms; and differences in age of onset of psychotic symptoms, age, severity of psychotic symptoms, severity of comorbid symptoms, risk of conversion to psychosis and functionality between OCS+ and OCS-patients.
One reviewer extracted individual study data onto a piloted extractions sheet. Another reviewer confirmed the extracted data. The following data items were collected, when available: Study design and country; inclusion and exclusion criteria; number of patients, average age and percentage female; type of population (FEP or at risk state); prevalence of OCD, prevalence of OCS, age of onset of psychotic symptoms, age of onset of OCS, age, Positive and Negative Syndrome Scale (PANSS) scores (Kay et al., 1987), rate of conversion to psychosis (in at risk patients only), scores in depression and anxiety scales, rate of suicide and Global Assessment of Funcion (GAF) scale scores.
The Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Studies Reporting Prevalence Data (Munn et al., 2015) was applied by one author. Adaptations to the scale were made when reasonable.
The principal summary measure was the mean of prevalences of OCD in FEP and in at risk patients. The mean of prevalences of OCS in FEP and in at risk patients was also calculated. Odds ratios were calculated to compare rate of conversion to psychosis and rate of suicide in OCS+ vs OCS-patients. Mean differences were calculated to compare age of onset of psychotic symptoms, age, PANSS+, PANSS-, depression, anxiety and GAF scores in OCS+ vs OCS-patients; and also to compare age of onset of psychotic symptoms with age of onset of OCS symptoms.
Data analyses were performed with JBI Sumari and SPSS (v 28.0.1.0). An estimate of prevalence of OCD and OCS and their respective confidence intervals were calculated in JBI Sumari using the Freeman-Tuckey transformation assuming a random-effects model. The the odds ratios and respective confidence intervals for the rate of conversion to psychosis and the rate of suicide in OCS+ and OCS-patients were calculated in JBI Sumari using an inverse variance method assuming a random-effects model. The mean differences and respective confidence intervals for the age of onset of psychotic symptoms, age, PANSS+, PANSS-, depression, anxiety and GAF scores in OCS+ and OCS-patients were calculated in SPSS assuming a random-effects model. Heterogeneity was quantified using the I 2 statistic. When possible, analyses were performed for two subgroups, FEP cohorts and at risk cohorts. Because different depression instruments were used, depression scores were standardised by calculating the percentage of each patient's score from the maximum score of the instrument, and subsequent statistical analyses were performed with this value.
Because previously anonymized published data was used, this research is exempt from institutional review board approval. The authors have no support or funding to report. The authors have no competing interests to report.

Results
Of the 243 articles screened, 37 were accepted for full-text assessment. Of the 11 articles identified from citation search, 5 were accepted for full-text assessment. Of those accepted for full-text assessment, 22 had any information on prevalence of OCD or OCS or on clinical characteristics of early psychosis patients who were OCS+ vs OCS-, and were included ( Fig. 1.). In Lencz et al. (2004), there were two cohorts, one of FEP patients and one of early psychosis patients. Articles were excluded if they reported on cohorts that had more complete or more recent data elsewhere, if they did not report on OCD or OCS prevalence or clinical characteristics of OCS+ vs OCS-patients, or if they were studies of only patients with early psychosis and OCD or OCS, since we were interested particularly in how OCS+ patients compare with OCS-patients. Two articles were excluded because several methodological issues arose when applying the JBI Critical Appraisal Checklist.
The JBI Checklist for Prevalence Studies was used to assess risk of bias in studies.
1. "Was the sample frame appropriate to address the target population?": All studies were marked "yes" because they gave a reasonable description of the sample frame that ensured its appropriateness to the target population. 2. "Were study participants recruited in an appropriate way? ": Seven studies were marked "unclear" because there was no mention of consecutive sampling (De Haan et al., 2013;Kennedy, 2021;Liu et al., 2022;McAusland, 2017;Niendam et al., 2009;Ong et al., 2021;Salokangas, 2016). All others were marked "yes". 3. "Was the sample size adequate?": All studies were marked "unclear" either because they were not specifically designed to estimate Records removed before screening: Duplicate records removed (n = 30)

Records screened (n = 213)
Records excluded (n = 151) Reports sought for retrieval (n = 37) Reports not retrieved (n = 0) Reports assessed for eligibility (n = 37) Reports excluded (n = 20): Other cohort more complete (n = 3) Did not report on any outcome (n = 11) All patients OCS or OCD (n = 5) Methodological issues (n = 1) Records identified from citation searching (n = 11) Reports assessed for eligibility (n = 11) Reports excluded (n = 6): Other cohort more complete (n = 3) Did not report on any outcome (n = 2) Methodological issues (n = 1)     prevalence or because if they were, no mention of sample size calculation was made. 4. "Were the study subjects and setting described in detail?": Ong et al.

Identification of studies via databases and registers
(2021) was marked "unclear" because its description of the setting was not as clear as other included studies. All others were marked "yes". 5. "Were the study subjects and setting described in detail?": Ong et al.
(2021) and Sim et al. (2006) were marked "yes". The remaining ones were marked "unclear" because there was no mention of analyses comparing patients that were included with patients that refused to participate.  Table 1, it was assumed that it was done using clinical judgment. In  it was mentioned that OCD was diagnosed using the SCID-I and it is suggested OCS were identified using clinical judgment; we decided to mark this study unclear because there was no mention of raters' experience, so there is no guarantee of the quality of clinical judgment. The remaining studies were marked "yes" either because when it is mentioned that OCD or OCS were assessed using clinical judgment, it is also mentioned that raters were experienced (Liu et al., 2022) or because there was mention of the use of scales (all others).
7. "Was the condition measured in a standard, reliable way for all participants?": Six studies (Choi et al., 2009;De Haan et al., 2013;Fontenelle et al., 2011;Hur et al., 2012;Ong et al., 2021;Salokangas, 2016; were marked "unclear" because there was no mention of the experience of raters and there was no mention of inter-rater reliability assessments. All other studies were marked "yes" because they fulfilled at least one of those conditions. 8. "Was there appropriate statistical analysis?": On 3 studies (DeVylder et al., 2012;Hui et al., 2013;Poyurovsky et al., 2008) this question is not relevant because they do not report on the prevalence of OCD or OCS and were included solely due to the report on clinical differences between OCS+ and OCS-patients. All other studies were marked "yes". 9. "Was the response rate adequate, and if not, was the low response rate managed appropriately?": All but three studies (Fontenelle et al., 2011;Kennedy, 2021;Liu et al., 2022) were marked "yes" either because they were cross-sectional (so did not have dropouts), or because drop-out rate was low, or because there was mention to a comparison of demographic characteristics between non-responders and responders and there were no differences.
Eighteen studies, totaling 19 cohorts and 3114 patients, included data on the prevalence of OCD, 12 relative to at risk patients and 7 relative to FEP patients. OCD was present in 8.9 % of patients (95 % CI  7.0 to 10.9 %, I 2 = 62.6 %). (Fig. 2). When only at risk studies were included, totaling 2357 patients, OCD was present in 7.9 % of patients (95 % CI 5.9 to 10.0 %, I 2 = 58.1 %). When only FEP studies were included, totaling 757 patients, OCD was present in 10.5 % of patients (95 % CI 8.3 to 12.8 %, I 2 = 53.7 %). Five studies included data on the prevalence of OCS, 3 relative to at risk patients and 2 relative to FEP patients, totaling 595 patients. OCS were present in 24. Three studies compared age of onset between OCS and psychotic symptoms, all of them relative to FEP patients, totaling 60 patients. Mean difference was − 0.49 years (95 % CI -1.74 to 0.77, p = 0.45, I 2 = 0 %).
Comparisons between OCS+ and OCS-patients may be found in Table 2. While there was a difference in PANSS positive subscale scores (OCS+ patients had a score 18.15 points higher than OCS-patients), this difference was not statistically significant. There were no statistically significant differences in age of onset, PANSS negative subscale score, risk of conversion to psychosis, scores of depression, HADS score, suicide rate and GAF score. Despite a statistically significant difference in age (OCS+ patients were younger) and in depression scores (OCS+ patients had higher depression scores) the differences in these metrics were minute. Since studies included in the depression score analysis reported data in different scales, scores were standardised to a 0-100 scale.

Discussion
We reached a prevalence of OCS of 21.4 % (8.3 to 38.2 %) in at risk patients and 34.0 % (26.3 to 42.1 %) in FEP patients; and a prevalence of   OCD of 7.9 % (5.9 to 10.0 %) in at risk patients and 10.5 % (8.3 to 12.8 %) in FEP patients. These figures are much larger than those found in the general population (3.5 % for OCD and 8.7 % for OCS) (Angst et al., 2004). Two previous studies reviewed the literature for the prevalence of OCS and OCD in at risk patients (Albert et al., 2018;, and found somewhat similar figures, but were not formally metaanalyzed. To our knowledge, there has not been any previous review of literature on the prevalence of OCD or OCS in FEP patients. The prevalence of OCS and OCD in schizophrenia have been estimated to be 30.7 % and 12.3 % respectively (Swets et al., 2014). Another metaanalysis reached 12.1 % for OCD in schizophrenia (Achim et al., 2011). There is substantial comorbidity between OCS and early psychotic states. For certain, the presence of OCS in these mostly antipsychoticnaïve patients, coupled with early 20th century reports of comorbidity before the use of antipsychotics, prove that there is a real co-occurrence of OCS and psychosis that goes beyond the iatrogenic effects of antipsychotics. However, these data do now allow to discern if this cooccurrence may represent two disorders that aggregate due to shared neurobiologic mechanisms or if there is a subset of psychotic patients who develop OCS secondary to psychotic symptoms.
There seems to be an increase in OCS and OCD prevalence as samples progress from healthy to at risk to FEP to established psychosis. This could be because OCS are associated with more severe psychosis. In fact, one of the longitudinal studies (Fontenelle et al., 2011) concluded that OCS prevalence was higher in the UHR sample than in the general population, but higher still in the UHR sample that would later convert to psychosis. We did find a trend association between OCS and PANSS+ scores.
We did not find an earlier age of onset for OCS relative to psychotic symptoms. This is in line with available evidence for schizophrenia (Devulapalli et al., 2008). However, our analysis included only four FEP studies totaling 60 patients. Also, psychosis onset in these studies (12.4, 15.1 and 23.4 years) was on the earlier end of the usual schizophrenia age of onset (Kessler et al., 2007), so these studies might not be representative of the usual FEP population.
When comparing OCS+ with OCS-patients, the most relevant positive finding was that OCS+ patients tended to have more severe positive symptoms in PANSS+, with an absolute difference of 18 points, that was not statistically significant. In a review of literature that included 18 schizophrenia studies (Cunill et al., 2009), OCS (but not OCD) were associated with greater positive and negative symptoms. We did not find any difference in PANSS-scores between OCS+ and OCS-patients. Given the frequent difficulty in assessing the nature of phenomena in psychiatry, we may wonder if this correlation of OCS with positive but not negative symptoms may actually represent a difficulty in distinguishing between obsessions, delusions and hallucinations. Several of the included studies stated that OCS and psychotic symptoms were assessed by experienced clinicians, and some highlighted a particular care in ensuring that OCS content was not related to psychotic content, so there is some reassurance that there was no phenomenological confusion in the assessment. However, others did not mention such care, and that could explain some of the heterogeneity of this analysis. These phenomenological discussions run deep in psychiatry and their importance shouldn't be disregarded here.
Since OCS may be associated with more severe psychotic illness, we expected a greater risk of progression to psychosis in the OCS+ at risk population. Despite that, we didn't find an increased risk, as the odds ratio was 0.82 (0.42 to 1.61). This is despite the fact that OCD increased the risk of psychosis in two large scale register studies (Meier et al., 2014, Cederlöf et al., 2015. These patients may not only represent a subsample of psychotic patients who also have OCS symptoms and therefore may have increased risk of progression, but may also represent a subset of OCD patients with low insight who also have psychotic symptoms and therefore don't have a significant risk of progression to psychosis.
OCS+ patients had marginally more severe depressive symptoms.
We could not find other instances in the literature where depressive symptoms in OCS+ vs OCS-psychotic patients were reviewed. This result is expected given the relation between OCD, depression and suicidality (Angelakis et al., 2015;Rasmussen and Tsuang, 1986) and between schizophrenia, depression and suicidality (Cassidy et al., 2018). On the other hand, we did not find an increased rate of suicide in OCS+ patients. However, we were only able to compare deaths by suicide and not rate of suicide ideation or suicide attempts. These measures could have been more sensitive to differences between the two groups. We found no differences in the prevalence of anxiety symptoms, which is surprising given the major comorbidity between OCD and anxiety. Nevertheless, only two studies including 56 patients were included in the analysis. Also unexpectedly, we did not find decreased functionality in OCS+ patients, which was described in previous literature for schizophrenia patients (Berman et al., 1995). This may be because the level of dysfunctionality in early stages of psychosis may be lower than that in chronic schizophrenia, so differences between groups may be harder to detect or nonexistent. In a meta-analysis that evaluated the contributing factors to decreased quality of life in FEP (Watson et al., 2018), one of the more notable correlations was with negative symptoms, which is interesting considering that we found no differences in negative symptom severity. It would be interesting to review other functionality measures as well as quality of life measures.
Several mechanisms may be responsible for comorbidity between OCS/OCD and early psychosis. Since prevalences of OCS and OCD are much higher in early psychosis relative to the general population, chance alone cannot be the sole explanation for comorbidity. Stengel, in his seminal 1944 article, regarded "the psychotic reaction as the result of a breakdown in the fight against odds which the patient had previously attempted to meet by the aid of neurotic defence mechanisms", so when patients with psychosis also had OCS, their "excessive inclination (...) to reality proving and doubt affected their attitude to psychotic experiences in a favourable way." It would be useful to know whether OCS decreased the risk of conversion to psychosis in at risk patients, in line with Stengel's point. We found neither a decreased nor an increased risk, however literature in this regard is too sparse to draw any conclusions. There may be shared neurobiologic factors in OCD and psychosis that explain highly prevalent comorbidity. OCS+ schizophrenia patients have more pronounced cognitive deficits that mirror partially overlapping neurobiological OCD mechanisms (Schirmbeck and Zink, 2013). Also, several genes have been found to be commonly associated with both OCD and schizophrenia (O'Connell et al., 2018). Both disorders suffer a significant effect from environmental factors both on onset and on severity. However the relationship between environmental factors and disease hasn't been studied in the context of comorbidity between OCS and psychosis. (Schirmbeck and Zink, 2013) The use of antipsychotic agents is a clear environmental cause of OCS (Schirmbeck et al., 2011), and may explain part of the excess prevalence in early psychosis patients who have already been treated with these agents. However, there is a significant part of patients whose OCS started before psychosis, thereby before the use of antipsychotics (Sterk et al., 2010), and a significant part of the patients reported in the studies included in this review had never been exposed to antipsychotics.
Current notions of obsession and compulsion are overinclusive. In patients with psychosis in particular, they overlap with psychotic experiences and self disorder manifestations (Rasmussen and Parnas, 2022). Confusion in the included studies regarding the precise nature of phenomena that were counted as obsessive or compulsive may also explain the high prevalence, and also the high heterogeneity. But most importantly this high prevalence points towards the need for careful phenomenological evaluation of experiences that may be proper comorbid OCS but may also be pseudo-obsessions or pseudo-compulsions that can be viewed as part of a wider concept of psychosis, and that distinction is not yet clear in the literature. In early psychosis states, where full-blown psychotic symptoms such as delusions or hallucinations are not yet clearly discernible, these other symptoms are of great relevance for diagnosis (Rasmussen and Parnas, 2022) and clear demarcation of subjective experiences is fundamental for treatment. Our results did not find a difference in age of onset of OCS and of psychosis, that could point the way in elucidating their relationship. However, this analysis included few patients and different clusters of patients, possibly with different types of phenomena, and more careful and insightful analyses are needed. A more clear homogeneity of clusters of patients will also be fundamental towards better research into treatment interventions. .
This study had several limitations. First, there was high heterogeneity in the reported data, which resulted in a low number of included studies in some of the comparisons. Second, there was high statistical heterogeneity in several of the analyses, namely the prevalence of OCS and the comparison of age of onset, PANSS scores and GAF scores. It should be noted that while the diagnosis method for OCD was fairly uniform among studies (10 in 17 used the DSM-IV), the assessment of OCS was much more variable, and even when studies used Y-BOCS, the criteria that were used for OCS+ classification varied greatly. Some studies only reported OCS that were unrelated to psychotic content and others reported on any OCS regardless of a possible psychotic nature. Also, in several studies the assessment of both OCS and OCD was made using clinical judgment, which can vary greatly from center to center. Third, one of the longitudinal studies (Fontenelle et al., 2011) identified high fluctuation in OCS throughout the prodrome, but our review included mostly cross-sectional studies, which means that subtleties in the evolution of OCS may have been lost. It must be said however that another longitudinal study (Niendam et al., 2009) found relative stability in OCS, so more studies are necessary to clarify their evolution. Fourth, since the number of included studies in each OCS+ vs OCSanalysis was low, we were not able to perform subgroup analysis for OCS and OCD samples and for at risk and FEP samples. That may also account for the high heterogeneity of the PANSS and the GAF comparisons. Fifth, we were only able to dichotomously compare OCS+ and OCS-patients, and not run analyses with OCS as a continuous variable. Sixth, we did not assess results in cognitive function, which may shed light on the results in functionality and risk of conversion that we obtained.

Conclusions
OCD was present in 7.4 % and 10.5 % and OCS in 21.4 % and 34.0 % of at risk and FEP patients respectively. Similar ages of onset for OCS and psychosis symptoms were found. OCS+ patients had statistically insignificant higher scores in the PANSS+ scale and had marginally higher depression scores. There were no differences between OCS+ and OCSpatients in age of onset, PANSS-score, risk of conversion to psychosis, anxiety score, suicide rate, and functionality scores.

Twitter
Obsessive-compulsive symptoms are frequent in psychosis. In this systematic review and meta-analysis, OCS are found to be very frequent in first episode psychosis and "at risk" states and may be associated with more severe psychosis.

Funding
No funding to report.

Declaration of competing interest
No conflicts of interest to report.