Significant outcomes
-
Cannabis cessation is associated with an increase of inflammation markers such as leucocytes, monocytes and lymphocytes in patients with psychosis.
-
Cannabis cessation is associated not only with an increase of inflammation markers such as leucocytes, monocytes and lymphocytes but also with hsCRP level in patients with schizophrenia.
-
Kinetics of monocyte levels are associated with symptomatology in patients with psychosis.
Limitations
-
The retrospective design of the cohort does not allow us to conclude a potential causal relationship.
-
The naturalistic design of the cohort does not evaluate other inflammatory markers than leukocytes, hsCRP and fibrinogen.
-
The naturalistic design does not permit to assess the impact of cannabis potency and its exact THC/cannabidiol ratio.
Introduction
A dysbalance of the immune system in psychotic disorders has been well investigated (Khandaker et al., Reference Khandaker, Cousins, Deakin, Lennox, Yolken and Jones2015). The vulnerability-stress-inflammation model (Howes & McCutcheon, Reference Howes and McCutcheon2017; Müller, Reference Müller2018; Lestra et al., Reference Lestra, Romeo, Martelli, Benyamina and Hamdani2022) implies an imbalance of the microglia activation (M1/M2 pathways’ homeostasis) leading to an over-expression of pro-inflammatory cytokines (Howes & McCutcheon, Reference Howes and McCutcheon2017). This imbalance could be due to microglia priming, including perinatal insults, early life stresses or genetic vulnerabilities, which leads to a pro-inflammatory M1 phenotype (Boksa, Reference Boksa2010; Sominsky et al., Reference Sominsky, Walker, Ong, Tynan, Walker and Hodgson2012; Diz-Chaves et al., Reference Diz-Chaves, Astiz, Bellini and Garcia-Segura2013; Hagberg et al., Reference Hagberg, Mallard, Ferriero, Vannucci, Levison, Vexler and Gressens2015) (5–8) and an overactivation/unregulated synaptic pruning especially in the prefrontal cortex and hippocampus, which may explain the negative and cognitive symptoms of schizophrenia (Boksa, Reference Boksa2010; Nestor et al., Reference Nestor, Kubicki, Nakamura, Niznikiewicz, Levitt, Shenton and McCarley2013; Sui et al., Reference Sui, Pearlson, Du, Yu, Jones, Chen, Jiang, Bustillo and Calhoun2015; Howes & McCutcheon, Reference Howes and McCutcheon2017). Moreover, this microglial overactivation could lead to an increased disinhibition of subcortical dopamine neurons, which is thought to underline the development of positive symptoms (Grace, Reference Grace2012; Kim et al., Reference Kim, Rossi, Aryal, Racz, Kim, Uezu, Wang, Wetsel, Weinberg, Yin and Soderling2015; Howes & McCutcheon, Reference Howes and McCutcheon2017).
The prevalence of cannabis consumption in patients with psychosis (more than 25%) is higher than in general population (Hunt et al., Reference Hunt, Large, Cleary, Lai and Saunders2018) but the bidirectional effect of cannabis use and psychosis has not yet been elucidated. Cannabis consumption leads to a poorer functional prognosis (Krause et al., Reference Krause, Huhn, Schneider-Thoma, Bighelli, Gutsmiedl and Leucht2019) and a higher rate of hospitalisations (Colizzi et al., Reference Colizzi, Burnett, Costa, De Agostini, Griffin and Bhattacharyya2018). Specific characteristics of this consumption seem to be associated with the risk of psychotic disorder, particularly notably for daily and high consumption (i.e. with high THC dosage among the components) (Marconi et al., Reference Marconi, Di Forti, Lewis, Murray and Vassos2016; Di Forti et al., Reference Di Forti, Quattrone, Freeman, Tripoli, Gayer-Anderson, Quigley, Rodriguez, Jongsma, Ferraro, La Cascia, La Barbera, Tarricone, Berardi, Szöke, Arango, Tortelli, Velthorst, Bernardo, Del-Ben and van der Ven2019). Indeed, there is a correlation between the intensity of cannabis consumption and the risk for psychosis (Marconi et al., Reference Marconi, Di Forti, Lewis, Murray and Vassos2016). The age of onset of cannabis use is also associated with an earlier onset of psychosis (Di Forti et al., Reference Di Forti, Sallis, Allegri, Trotta, Ferraro, Stilo, Marconi, La Cascia, Reis Marques, Pariante, Dazzan, Mondelli, Paparelli, Kolliakou, Prata, Gaughran, David, Morgan, Stahl, Khondoker, MacCabe and Murray2014; Setién-Suero et al., Reference Setién-Suero, Martínez-García, de la Foz, Vázquez-Bourgon, Correa-Ghisays, Ferro, Crespo-Facorro and Ayesa-Arriola2018) and an increase in psychotic disorder risk (Di Forti et al., Reference Di Forti, Quattrone, Freeman, Tripoli, Gayer-Anderson, Quigley, Rodriguez, Jongsma, Ferraro, La Cascia, La Barbera, Tarricone, Berardi, Szöke, Arango, Tortelli, Velthorst, Bernardo, Del-Ben and van der Ven2019).
White blood cells (WBC) count, an indirect marker of low-grade inflammation, has been associated with schizophrenia (Jackson & Miller, Reference Jackson and Miller2020; Mazza et al., Reference Mazza, Lucchi, Rossetti and Clerici2020). Schizophrenia and first-episode psychosis are associated not only with highest levels of blood total WBC, monocytes and neutrophils than controls (Jackson & Miller, Reference Jackson and Miller2020) but also higher neutrophils/lymphocytes ratio and monocytes/lymphocytes ratio (Mazza et al., Reference Mazza, Lucchi, Rossetti and Clerici2020). Moreover, specific features of monocyte cell functioning seem to be associated with schizophrenia. This includes higher gene expression of inflammatory components (Drexhage et al., Reference Drexhage, van der Heul-Nieuwenhuijsen, Padmos, van Beveren, Cohen, Versnel, Nolen and Drexhage2010), increased production of IL1b (Uranova et al., Reference Uranova, Bonartsev, Androsova, Rakhmanova and Kaleda2017), and intracellular IL6 (Krause et al., Reference Krause, Wagner, Wildenauer, Matz, Weidinger, Riedel, Obermeier, Gruber, Schwarz and Müller2012).
Few studies have investigated the impact of cannabis consumption on inflammatory markers with psychiatric disorders. Recently, we have found in a small sample, that cannabis cessation increases inflammatory markers such as high sensitivity C-reactive protein (hsCRP) and lymphocyte count, as compared to non-cannabis users, cannabis users with schizophrenia display lower levels of inflammation (Romeo et al., Reference Romeo, Lestra, Martelli, Benyamina and Hamdani2022). These observations were in line with other studies (Miller et al., Reference Miller, Buckley and McEvoy2018; Gibson et al., Reference Gibson, Bassir Nia, Spriggs, DeFrancisco, Swift, Perkel, Zhong, Mazumdar, Fernandez, Patel, Kim-Schulze and Hurd2020) although with discrepancy. For example, when using a within-subject design also in a small sample, Goetz and Miller (Reference Goetz and Miller2019) did not find any difference in total nor differential WBC levels between acutely ill patients with schizophrenia with and without cannabis use.
Given the paucity of data as well as the high prevalence of THC use in patients with psychosis, further exploration of the relationship between THC use and inflammatory markers in this population is necessary. Therefore, the objective of this retrospective study was to investigate the impact of cannabis use on inflammatory markers (WBC, hsCRP and fibrinogen levels) in a larger sample of patients with psychosis and to explore the link between these inflammatory markers and clinical symptoms.
Methods
Participants
One hundred and two inpatients were included in this retrospective study from the psychiatry and addictology department at Paul Brousse Hospital (Paris). The inclusion criteria were (i) inpatients in an acute phase of psychosis (including schizophrenia, schizoaffective disorder, first episode of psychosis (FEP), non-specific delusional disorders, bipolar disorder and major depressive disorder with psychotic symptoms); (ii) over 18 years old and (iii) available urinary THC data on admission. Exclusion criteria included neurological disorders, intellectual disability, acute infection or auto-immune pathologies, high levels of inflammatory markers at baseline (hsCRP level >20 mg/mL and/or WBC >10 Giga/L) and use of anti-inflammatory treatments. In this non-interventional retrospective cohort study only using existing data from routine care, no supplementary clinical or biological examinations were done and so this study did not change the standard of care provided to the patients. A complete information was given to all patients on the possible retrospective use of their routine care data for research purposes. All patients were informed that they could refuse to have their data used retrospectively without impacting or changing the care provided. None of the included patients refused. The study was conducted in accordance with the Declaration of Helsinki and was approved by a French national committee (N°1980120).
Data collected
Socio-demographic data, comorbidities, undergoing treatments, smoking status and body mass index (BMI) were collected. Positive and Negative Syndrome Scale (PANSS) score was assessed by a trained psychiatrist. Lipid profile, fasting blood sugar, fibrinogen, hsCRP and WBC results were collected at baseline and 4 weeks later. Urinary cannabis was also collected on admission and 4 weeks later using an immuno-enzymatic method (C800 Abbott ®).To evaluate the impact of cannabis cessation on inflammatory markers, in the THC+ group, only patients who had divided at least by two their baseline THC levels at 4 weeks and with a THC level <100 ng/mL were included.
Data analyses
Analyses were performed with Medistica.pvalue.io, a graphic user to the R statistical analysis software for scientific medical publications (Available on: https://www.pvalue.io). The normal distribution of the different data sets was evaluated by a Shapiro–Wilk test. For samples including more than 30 patients with normal distribution of parameters, a t-test was used to compare patients in the THC+ group versus THC− group. If we had a smaller number of patients (<30 patients) and/or the data distribution was not normal, we used a Mann–Whitney non-parametric test. For categorical variables, a Chi-square or Fisher test was used. To assess the association between inflammatory markers and socio-demographic or PANSS score, Spearman or Pearson correlations were computed. A multivariate analysis was performed by linear regression including confounding factors such as age, gender, BMI, smoking statusand diagnosis.
Because more than 50 patients with schizophrenia were included, a subgroup analysis was also performed including only these patients. Given the small number of our sample, we chose to perform two models for the multivariate analysis: (i) the first one including available factors that are well known to be pro-inflammatory: smoking status and BMI as covariates and (ii) the second including patients’ main intrinsic factors: gender and age as covariates.
Results
Psychosis patients
Description of the population
One hundred and two patients with psychosis were included in this study (65 THC− vs 37 THC+) at baseline and 68 attended follow-up at 4 weeks (48 THC− vs 20 THC+). The main characteristics of this population are described in Table 1. BMI was higher in the THC− group (24.7 vs 22.2; p < 0.01), patients were older in the THC− group (40.4 vs 32.1 years; p < 0.001). A lower proportion of tobacco smokers was found in the THC− group (38% vs 95%; p < 0.001). The diagnosis repartition was also different between the two groups (p < 0.01). We did not find any difference between the two groups regarding gender, PANSS total score and subscores. The groups did not differ in the type of treatment at baseline.
Table 1. Baseline characteristics of patients with psychosis
SCZ: schizophrenia, SCZ aff: schizoaffective disorder, FEP: First episode psychosis, MDD: Major depressive disorders, BD: Bipolar disorder, THC: tetrahydrocannabinol, BMI: Body Mass Index, AP: antipsychotic, AD: antidepressant, MS: mood stabiliser, BZD: Benzodiazepine, LDL: Low-density lipoprotein cholesterol, HDL: High-density lipoprotein cholesterol.
THC and markers of inflammation
At baseline
We found highest total WBC (p = 0.04), lymphocyte (p = 0.01) and monocyte (p = 0.01) counts in the THC + group. No difference was found neither for neutrophil, basophil, eosinophil and platelet counts, nor for fibrinogen level or the proportion of patients with CRP level above 3 mg/L. When considering confounding factors such as age, gender, smoking status, BMI and diagnosis, the results did not persist (Table 2 and Fig. 1A).
Fig. 1. A. Variations of leucocyte, lymphocyte, monocyte levels in patients with psychosis THC + and THC− between baseline and at 4 weeks. No difference was found between THC + and THC− groups at baseline for leucocyte, lymphocyte, monocyte levels. Higher levels were found in THC + for leucocyte, lymphocyte and monocyte levels, than in THC− at 4 weeks. Significant increase between baseline and 4 weeks were found in THC + group for leucocyte, lymphocyte and monocyte levels. Significant increase was found between baseline and 4 weeks in THC− group for leucocyte levels. B. Variations of leucocyte, lymphocyte, monocyte levels in patients with schizophrenia THC + and THC− between baseline and at 4 weeks. A lower level of monocytes was found in the SCZ THC− group as compared to the SCZ THC + group at baseline. Higher levels were found in SCZ THC + than in SCZ THC− at 4 weeks for leucocytes, lymphocytes and monocytes. Significant increase between baseline and 4 weeks were found in SCZ THC + group for leucocyte, monocyte levels and a statistical trend to an increase was found for lymphocyte level. Significant increase was found between baseline and 4 weeks in SCZ THC− group for leucocyte levels. * p < 0.05; # p = 0.06.
Table 2. Summary of psychosis results
Adjusted for BMI, tobacco status, diagnosis and age.
At 4 weeks
We found highest WBC (p = 0.02), lymphocyte (p < 0.01) and monocyte (p = 0.01) counts in the THC + group. No difference was found neither for neutrophil, basophil, eosinophil and platelet counts, nor for fibrinogen level or the proportion of patients with CRP level above 3 mg/L. When considering confounding factors such as age, gender, smoking status, BMI and diagnosis, all the differences found in univariate analysis remained significant, respectively, for WBC (p = 0.03), lymphocyte (p = 0.04) and monocyte (p < 0.01) counts. Between baseline and 4 weeks, we found a significant increase of WBC (p < 0.01), lymphocyte (p = 0.01), monocyte (p < 0.01) eosinophil (p < 0.01) and neutrophil (p = 0.04) counts in the THC + group and a significant increase of WBC (p = 0.01), eosinophils (p < 0.01) and platelets (p = 0.04) in the THC− group (Table 2 and Fig. 1A). This increase of leucocyte level (p = 0.01), lymphocyte level (p < 0.01) and monocyte level (p = 0.03) between baseline and 4 weeks follow-up was greater in the THC + group than in the THC− group. These differences remained significant when accounting for confounding factors such age, gender, smoking status, BMI and diagnosis for leucocyte level (p < 0.01) and monocytes (p = 0.05), with persistent statistical trends for lymphocyte (p = 0.08) levels (Table 2).
Correlation between baseline inflammatory markers and clinical variables
In the THC + group at baseline, we found a negative correlation between lymphocytes count and PANSS negative subscore (r = −0.383; p = 0.04). We found a negative correlation between fibrinogen level and PANSS general subscore at baseline (r = −0.421; p = 0.036) or at 4 weeks (r = −0.537; p = 0.02) and with PANSS total (r = −0.537; p = 0.02) or PANSS positive subscore (r = −0.618; p < 0.01) at 4 weeks.
Correlation between inflammatory markers at 4 weeks and clinical variables
In the THC + group, we found a negative correlation between lymphocyte count and baseline PANSS general subscore (r = −0.497; p = 0.03). We found a positive correlation between monocyte count and baseline PANSS negative subscore (r = 0.452; p = 0.045). We found a statistical trend for a positive correlation between the variation of monocyte count between baseline and 4 weeks and the PANSS total score at 4 weeks (r = 0.443; p = 0.05), the PANSS general subscore at 4 weeks (r = 0.393; p = 0.09) and the variation of PANSS general subscore (r = 0.416; p = 0.07).
Patients with schizophrenia
Description of the population
Sixty-four patients with schizophrenia were included in this study (49 in SCZ THC− vs 25 in SCZ THC+) at baseline and 45 attended follow-up at 4 weeks (33 in SCZ THC- vs 12 SCZ THC+). The main characteristics of this sub-population are described in Table 3. No difference was found between the two groups in terms of age, gender, BMI, illness duration, baseline medication, diagnosis or baseline PANSS total score and subscores. A higher proportion of tobacco smokers was found in the SCZ THC + group (92% vs 44%; p < 0.01) (Table 3).
Table 3. Baseline characteristics of patients with schizophrenia
SCZ: schizophrenia, SCZ aff: schizoaffective disorder, THC : tetrahydrocannabinol, BMI: body mass index, AP: antipsychotic, AD: antidepressant, MS: mood stabiliser, BZD: benzodiazepine, LDL: Low-density lipoprotein cholesterol, HDL: High-density lipoprotein cholesterol.
THC and markers of inflammation
At baseline
We found highest WBC (p = 0.024), lymphocyte (p = 0.021) and monocyte (p = 0.035) counts in the SCZ THC + group compared to the SCZ THC− group. These differences did not remain significant when we included confounding factors such as tobacco status and BMI, except for monocyte count (p = 0.042) and for WBC when we included confounding factors such as age and gender (Table 4 and Fig. 1B). No difference was found between these two groups for fibrinogen, neutrophil, basophil, eosinophil and platelet counts. We also found a higher proportion of patients with a CRP above 3mg/L in the SCZ THC− (p = 0.043).
Table 4. Summary of schizophrenia results
Adjusted for * BMI, tobacco status; ** Age, gender.
At 4 weeks
We found highest WBC (p = 0.03), lymphocyte (p = 0.01) and monocyte (p < 0.01) counts in the SCZ THC + group. These differences remained significant after adjustment for smoking status and BMI only for WBC (p = 0.04), lymphocyte (p = 0.03) and monocyte (p < 0.01) counts, and after adjustment for age and gender for WBC (p = 0.02), lymphocyte (<0.01) and monocyte (<0.01) counts. No difference was found neither for neutrophil, basophil, eosinophil and platelet counts nor for fibrinogen level or the proportion of patients with CRP above 3 mg/L. In the SCZ THC + group, between baseline and 4 weeks, we found a significant increase in WBC (p < 0.01), monocyte (p < 0.01), eosinophil (p < 0.01) counts and a statistical trend for lymphocyte (p = 0.055), basophil (p = 0.068) and neutrophil (p = 0.092) counts. In the SCZ THC− group, we found a significant increase in WBC (p = 0.045), basophil (p = 0.01), eosinophil (p < 0.001) counts and a statistical trend for platelet (p = 0.07) count (Table 4 and Fig. 1B). Higher increase of leucocyte level (p = 0.03), lymphocyte level (p = 0.04) and monocyte level (p < 0.01) were found in THC + group than in THC− group after 4 weeks. These differences remained significant when accounting for confounding factors such as smoking status and BMI for leucocyte (p = 0.01) and monocyte (p = 0.02) levels and statistical trends were found for lymphocyte level (p = 0.09). When accounting for confounding factors such as age and gender, leucocytes (p = 0.01), monocytes (p = 0.03) continued to be significantly higher in the SCZ THC + group at 4 weeks, and statistical trends were found for lymphocyte levels (p = 0.06) (Table 4).
Correlations between inflammation markers and clinical variables
In the SCZ THC + group, we found negative correlations between monocyte level at 4 weeks and the variation PANSS positive subscore (r = −0.579; p = 0.05) and PANSS total score (r = −0.606; p = 0.037) between baseline and 4 weeks.
Discussion
In this study, we found kinetic differences of WBC, lymphocyte and monocyte levels between patients with psychosis consumers who stopped using cannabis at 4 weeks as compared to non-consumers. Indeed, increased levels of WBCs, monocytes and lymphocytes were found at 4 weeks in patients who stopped using cannabis. We also found similar results in a subgroup of patients with schizophrenia. Furthermore, we found that monocyte levels were associated with PANSS-negative symptoms in patients with psychosis among the THC + group.
The impact of cannabis use on inflammatory markers in patients with psychiatric disorders has been evaluated in few studies (24–28) (Miller et al., Reference Miller, Buckley and McEvoy2018; Goetz & Miller, Reference Goetz and Miller2019; Gibson et al., Reference Gibson, Bassir Nia, Spriggs, DeFrancisco, Swift, Perkel, Zhong, Mazumdar, Fernandez, Patel, Kim-Schulze and Hurd2020; Corsi-Zuelli et al., Reference Corsi-Zuelli, Marques, da Roza, Loureiro, Shuhama, Di Forti, Menezes, Louzada-Junior and Del-Ben2021; Romeo et al., Reference Romeo, Lestra, Martelli, Benyamina and Hamdani2022) (25–29). Levels of specific inflammatory markers such as IL-6, IFN-G and CRP were found to be lower among cannabis users as compared to non-users (Miller et al., Reference Miller, Buckley and McEvoy2018; Gibson et al., Reference Gibson, Bassir Nia, Spriggs, DeFrancisco, Swift, Perkel, Zhong, Mazumdar, Fernandez, Patel, Kim-Schulze and Hurd2020). Goetz and Miller (Reference Goetz and Miller2019) examined the effect of marijuana use on WBC counts in acutely ill inpatients with schizophrenia, using both within-subjects and between-groups designs. They did not find any significant impact of cannabis use on these markers. These results are in line with ours, as we did not find any difference at baseline between THC users and non-users. To our knowledge, our latest work is the first follow-up of selected inflammatory markers after cannabis cessation in patients with schizophrenia (Romeo et al., Reference Romeo, Lestra, Martelli, Benyamina and Hamdani2022). In this first study, we have shown that CRP levels were lower at baseline in the SCZ THC + group, whereas no difference was observed at 4 weeks between the two groups, suggesting a restoration of low-grade inflammation after cannabis cessation. In this study, including a larger population, we confirmed these data with lowest levels of CRP in the SCZ THC + group as compared to the SCZ THC− group at baseline, and no significant difference after 4 weeks of follow-up considering the CRP levels. In our first work, we also found that lymphocyte levels were higher both at baseline and after 4 weeks of follow-up in the SCZ THC + group, with a significant increase across study period at baseline (Romeo et al., Reference Romeo, Lestra, Martelli, Benyamina and Hamdani2022). We did not find the same result in this study with no difference at baseline between the THC + and THC− group. The difference in sample sizes between the two studies might explain these different results (Romeo et al., Reference Romeo, Lestra, Martelli, Benyamina and Hamdani2022).
The higher level of WBCs, lymphocytes and monocytes at 4 weeks and between baseline and 4 weeks found in patients in psychosis, notably in schizophrenia with cannabis use, reinforces the idea that cannabis use acts on peripheral and central inflammation (Cabral et al., Reference Cabral, Rogers and Lichtman2015). The interaction between THC and CB2 receptors expressed on lymphocyte and monocyte membranes could explain the modulation of these cells’ migration and cytokine production (Sexton et al., Reference Sexton, Silvestroni, Möller and Stella2013; Persidsky et al., Reference Persidsky, Fan, Dykstra, Reichenbach, Rom and Ramirez2015). This interaction could explain the decrease in microglial activity, known to be increased in psychosis (Pollak et al., Reference Pollak, Beck, Irani, Howes, David and McGuire2016). Another hypothesis that may explain the modulation of inflammatory markers by cannabis use could be its impact on the blood–brain barrier (BBB) and intestinal permeability. Indeed, cannabis use has been associated with decreased permeability and decreased monocyte translocation across the BBB in patients with human immunodeficiency virus (Persidsky et al., Reference Persidsky, Fan, Dykstra, Reichenbach, Rom and Ramirez2015; Ellis et al., Reference Ellis, Peterson, Li, Schrier, Iudicello, Letendre, Morgan, Tang, Grant and Cherner2020). Besides, abnormalities in BBB permeability and its impact on central translocation of pro-inflammatory cytokines are known to be linked with increased intracerebral inflammation in patients with schizophrenia (Pollak et al., Reference Pollak, Drndarski, Stone, David, McGuire and Abbott2018). Cannabis may affect the intestinal permeability, as observed in patients with psychosis (Ishida et al., Reference Ishida, Ogura, Aizawa, Ota, Hidese, Yomogida, Matsuo, Yoshida and Kunugi2022). Indeed, THC decreases intestinal permeability (36–39) (Massa et al., Reference Massa, Marsicano, Hermann, Cannich, Monory, Cravatt, Ferri, Sibaev, Storr and Lutz2004; Kimball et al., Reference Kimball, Schneider, Wallace and Hornby2006; Alhamoruni et al., Reference Alhamoruni, Lee, Wright, Larvin and O’Sullivan2010; Massa et al., Reference Massa, Marsicano, Hermann, Cannich, Monory, Cravatt, Ferri, Sibaev, Storr and Lutz2004) (37–40) through the tight junction proteins (Alhamoruni et al., Reference Alhamoruni, Lee, Wright, Larvin and O’Sullivan2010). The increase in inflammatory markers after 4 weeks of cannabis cessation could also point to a resumption of BBB and gut abnormal permeability.
We also found in the THC + group that monocyte level and monocyte change across study period were correlated with the PANSS score. THC use in patients with psychosis is associated with poorer functionality (Setién-Suero et al., Reference Setién-Suero, Martínez-García, de la Foz, Vázquez-Bourgon, Correa-Ghisays, Ferro, Crespo-Facorro and Ayesa-Arriola2018). An association between poor functioning, negative and neurocognitive symptoms in patients with psychosis has been described (Stouten et al., Reference Stouten, Veling, Laan, van der Helm and van der Gaag2017; Lee et al., Reference Lee, Hermens, Naismith, Kaur, Guastella, Glozier, Scott, Scott and Hickie2018). Interestingly, cognitive symptoms are associated with inflammatory markers in somatic and psychiatric diseases (Fourrier et al., Reference Fourrier, Singhal and Baune2019) which could be explained by a dysregulation of the glutamatergic, tryptophan-kynurenine pathway and hypothalamic-pituitary axis (Fourrier et al., Reference Fourrier, Singhal and Baune2019). Minocycline by reducing microglial activation has been proposed to improve cognitive impairment among schizophrenic patients (Levkovitz et al., Reference Levkovitz, Mendlovich, Riwkes, Braw, Levkovitch-Verbin, Gal, Fennig, Treves and Kron2010). We can suspect that increased inflammation in patients with psychosis could lead to a neurocognitive deterioration as observed in HIV patients (Williams et al., Reference Williams, Veenstra, Gaskill, Morgello, Calderon and Berman2014; Hong & Banks, Reference Hong and Banks2015). Cannabis use could therefore be used to counteract anxiety symptoms and therefore cognitive impairment (Schofield et al., Reference Schofield, Tennant, Nash, Degenhardt, Cornish, Hobbs and Brennan2006; Schaub et al., Reference Schaub, Fanghaenel and Stohler2008; Dekker et al., Reference Dekker, Linszen and De Haan2009). The withdrawal period is well known by clinicians as a time of high vulnerability, in particular due to the anxiety symptoms which can trigger acute relapses. Taking into account the increase in inflammation during this period could therefore open therapeutic perspectives during cannabis withdrawal for these patients.
Our study has several limitations. First, the retrospective design of the cohort does not allow us to conclude for a potential causal relationship. Furthermore, because of the retrospective design of our study, we were enabled to assess the addictological profile of the patients which could bias our data such as tobacco/cannabis, notably we did not have information on the frequency of cannabis use, the cannabis potency nor the exact THC/cannabidiol ratio. Finally, we did not study other inflammatory markers such as cytokine levels. These three limitations can be explained by the naturalistic, non-interventional design of our study using existing data from routine care and need to be replicated by other studies.
In conclusion, this study shows that cannabis cessation is associated with an increased inflammation depicted by an elevation of white blood cell, lymphocyte and monocyte levels, which correlates with symptomatology of patients with psychosis. Studying the link between cannabis and inflammation could lead to a better understanding of the pathophysiology of psychosis.
Acknowledgements
None.
Authors contributions
Dr Romeo, V. Lestra, Pr Benymina and Dr Hamdani designed the present study. Dr Romeo, V. Lestra and Dr Martelli managed the acquisition of data. Dr Romeo undertook the statistical data analysis. Dr Romeo and V. Lestra wrote the first draft of the manuscript. Dr Hamdani, Dr Martelli and Pr. Benyamina oversaw its revision. All authors contributed to and have approved the final manuscript.
Financial support
None.
Statement of interest
Bruno Romeo, Valentine Lestra, Catherine Martelli and Nora Hamdani have no conflict of interest. Amine Benyamina has given talks for Lundbeck, Mylan, Merck-Serono and Bristol-Myers Squibb and member of bord Indivior.
Ethical statement
The study was conducted in accordance with the Declaration of Helsinki and was approved by a French national committee (N°1980120).
