As for solids, the process of manufacturing a pure CBD isolate usually entails a distillation or crystallisation of sorts, which would remove metal residues to an extent. When considering the more stringent inhalation USP/ICH limits, presented in Fig. 2, the three metals that were responsible for most failures were mercury , nickel  and lead . Furthermore, only three of the categories were considered for the inhalation limits since products from the other categories had none or low potential for being dosed in an inhalation form. It is evident that a clear increase in the number of residue failures is present when applying the more stringent inhalation limit. A 346% increase is observed in the number of failures detected when inhalation limit residues are compared to oral limit residues. Note that 46%  of all failures are attributed to Class 1 residues, which is lower than the samples submitted for comparison with the oral limit. It should however be made clear that 87%  of failures are attributed to a combination of Class 1 and Class 2A. Next to Class 1, Class 2A residues are the second most dangerous to human health. For samples submitted to be tested against the inhalation specification limit, a significant number were vape pens or vape pen attachments containing oil. These pens or attachments were manufactured with a transparent glass exterior to be able to view the level of oil, which is contained in a metal housing. Observations made suggests that the metal housing showed characteristics of nickel plating. This could explain the vast number of nickel failures detected, since nickel is most probably leached into the product after production in finished form. The small quantity oil combined with the large surface area of the metal housing could leach enough nickel into products to cause detrimental health effects when inhaled. If no limit is enforced and only occurrence or presence of a residue in a sample is considered , four main residues make up 52%  of all residue occurrences. These residues include lead , cobalt , nickel  and thallium . What is more concerning is that a Class 1 residue was detected in 36%  of the samples. This means that, among all residues analysed, a Class 1 residues will be present in a third of all samples. Limits of detection are shown in both Table A3 and Table A5. Considering the sample failures presented in Fig. 1 pie chart, a total 15%  of all samples submitted for heavy metal residue analysis failed when compared against the USP/ICH oral specification limit. These samples included a wide variety of product types; drops, capsules, drinks, edibles and suppositories.

The latter was also considered under the oral specification limit although they are not exposed to low pH stomach acids which could decarboxylate certain cannabinoids. A staggering increase in the number of failures for the ICH inhalation limit is evident compared to the oral limit presented in Fig. 2 pie chart. A total of 44%  of all samples, in the applicable categories, submitted for heavy metal residue analysis failed when compared to the inhalation limit. This is a concern since almost half of all samples grouped in these categories failed. Sample types in this category included vape pens, raw plant material and vape oils. Furthermore, when considering the individual residues, failures are attributed mostly to Class 1 residues when applying the oral limit, and Class 1 as well as Class 2A residues when comparing to the inhalation limit. Class 1 and Class 2A heavy metal residues are considered to be most detrimental to human health. 88%  of samples contained heavy metal residues, irrespective of whether they are above or below specification limits as seen in Fig. 3 pie chart, this data may be used to focus on the metal residues most predominantly found in the analysed matrices. Chronic, as opposed to acute inflammation, has been proposed to be involved in the pathophysiology of several chronic physical health conditions, including coronary artery disease, diabetes, cancer, Alzheimer’s, osteoarthritis, and autoimmune diseases . In 2017, these four chronic conditions accounted for nearly 1.5 million deaths in the United States  . Additionally, a growing body of new evidence suggests that chronic low-grade inflammation is involved in mental health conditions and recurrence such as depression , as well as chronic pain . As such, patients and clinicians, are increasingly exploring novel therapies for chronic health conditions that target or mitigate chronic inflammation . The active constituents of cannabis, particularly tetrahydrocannabinol  and cannabidiol , have been shown to have immunomodulatory effects , specifically anti-inflammatory properties . In non-human primates, THC administration attenuated tissue inflammation . Consequently, there has been an increase in cannabis square pot use for medical purposes , particularly among conditions with an inflammatory component including HIV, cancer and chronic pain. Currently, 33 states have laws allowing medical cannabis use for a wide range of conditions, with 11 states allowing recreational cannabis use . Yet, there have not been many studies evaluating the association between cannabis use and inflammation in humans and findings from the few studies published have been mixed. Cannabis use was significantly associated with lower levels of C reactive protein , but only among those whose CRP levels were below the median . Lifetime but not recent cannabis use was associated with lower levels of fibrinogen. Further, studies have not found any significant association between cannabis use and interleukin 6  and CRP, high sensitivity CRP.

Furthermore, cumulative cannabis use  and cannabis dependence was not associated with levels of CRP.In contrast, any cannabis was significantly associated with elevated levels of CRP and lifetime cannabis use was significantly associated with lower levels of IL-6 . We aim to evaluate three biomarkers of systemic inflammation including IL-6, fibrinogen and hs-CRP – a more sensitive CRP test that can detect small changes in CRP levels , which may be more apt to estimating cannabis-inflammation effect. Furthermore, to understand the cannabis-inflammation association better, studies need to address for confounding variables particularly sociodemographic, behavioral and pharmacological confounders. Finally, studies on a potential sex-dependent effect of cannabis use on inflammation is scarce. In other studies, women reported higher ratings of subjective effects of cannabis use compared to men , while men exhibited greater cannabis-induced analgesia compared to women . Women are more likely to report loss of appetite, while men are more likely to report increased appetite associated with cannabis use . The mechanisms underlying potential sex-dependent effects of cannabis use may be attributed to sex-dependent differences in cannabis metabolism and interactions between the endocannabinoid system and sex hormones . As such, the evidence suggests that women are more sensitive than males to the behavioral and physiological effects of cannabis. This emerging evidence on potential sex-dependent effects of cannabis call for investigating cannabis use by sex interactions on health, including chronic inflammation. Therefore, the objective of this analysis was the examine the relationship between self-reported cannabis use and biomarkers of systemic inflammation, specifically high-sensitivity C-reactive protein , Interleukin 6  and fibrinogen. In light of findings from the extant literature , we hypothesized that self-reported cannabis use will be associated with significantly lower levels of each biomarker of systemic inflammation compared to nonuse. Additionally, we investigated self-reported cannabis use by sex interactions to determine whether sex-dependent differences emerge on the impact of self-reported of cannabis use on biomarkers of systemic inflammation.Data came from Wave 1 of the Population Assessment of Tobacco and Health  study Biomarker Restricted-Use files. The PATH study is a nationally representative, longitudinal cohort study of tobacco use and health outcomes in the United States . The PATH study is a collaboration between the National Institute on Drug Abuse , National Institutes of Health , the Center for Tobacco Products , and the Food and Drug Administration  to collect data on tobacco use patterns and related health outcomes from non-institutionalized residents of the U.S. aged 12 years and older.

Extensive details of the PATH study design and methods has been published previously . Of the 32,320 respondents who completed the Wave 1 adult interview, 21,801  provided a urine specimen. Among these, a stratified probability sample of 11,522 adults were selected from a diverse mix of six tobacco product use groups including:  current exclusive established users of cigarettes , current established users of one or more tobacco products other than cigarettes,current experimental users of any tobacco products,  former established users of any tobacco product ,  never users of any tobacco products and  current established users of cigarettes who are experimental users of at least one other tobacco products. Of the 11,522 adults, 7,159 also provided a blood specimen. Blood was collected from consenting adults at a separate study visit by a phlebotomist, who visited the respondent’s home. Blood specimens were shipped to the Division of Laboratory Sciences, National Center for Environmental Health , Centers for Disease Control and Prevention  for analysis of biomarkers of inflammation. Among the 7,159 who provided a blood specimen, we excluded respondents who reported that a doctor told them that they had a heart disease or cancer  Thus, the analysis sample comprised of 5,363 adults of PATH Wave 1.We used weighted frequencies, percentages , medians, and inter-quartile range  to describe the characteristics of the Wave 1 PATH sample by their self-reported cannabis use. The primary predictor in this analysis was the four-level variable of self-reported cannabis use and the primary outcomes were the four biomarkers of inflammation . All the outcome variables were highly skewed and were subsequently natural log-transformed to stabilize their distributions.We conducted the crude and adjusted models. We conducted the adjusted models in stages, model 1 adjusted for age and sex, model 2 additionally adjusted for race/ethnicity status, educational attainment, anti-inflammatory medication use, recency of alcohol use, recency of tobacco smoking and any illicit drug use in past 12 months, model 3 additionally adjusted for BMI. The set of covariates included in the final models were based on a priori knowledge of their relationship to biomarkers of systemic inflammation . To explore potential sex differences in association between self-reported cannabis use and the biomarkers of systemic inflammation,trim tray we repeated all analysis above and included a self-reported cannabis use by sex interaction term in the fully adjusted model . The analysis was conducted in SAS version 9.4  and accounted for the PATH study’s multi-stage stratified area probability sampling design and non-response adjustments by using Wave 1 blood biomarker weights , balanced repeated replication  and Fay’s adjustment set to 0.3 to increase estimate stability .

The replication weights is also recommended for sub-population analysis .Distribution of the untransformed biomarkers of systemic inflammation in the total population and by cannabis use groups and relationships among the biomarkers of inflammation are included in the Supplementary Material Tables S1& S2. Fig. 1 displays mean estimated hsCRP, IL-6 and fibrinogen levels  among categories of cannabis use. These results show a general pattern of lower mean biomarkers of systemic inflammation among respondents self-reporting cannabis use in the past 30 days compared to other categories of cannabis use. We compared each category of cannabis use with the group that self-reported never use. hsCRP. In bivariate analysis, cannabis use within the past 30 days was associated with lower levels  of hs-CRP compared to nonuse, with narrow confidence intervals around the beta estimates that excluded null and positive values . Adjusting for covariates in model 2 attenuated the beta estimate of this association, with confidence intervals also excluding null and positive values. In model 3 that additionally adjusted for BMI, the beta estimate was further attenuated, indicating lower levels of hsCRP in respondents who self-reported past 30 days cannabis use compared to nonuse, with confidence intervals around the point estimates excluding null value. All other categories of self-reported cannabis use indicated higher levels of hsCRP compared to nonuse , but the confidence intervals around the points estimates spanned both the null and negative values. IL-6. In bivariate analysis, self-reported cannabis use within the past 30 days was associated with lower levels of IL-6 compared to nonuse, with confidence intervals around the beta point estimates excluding null and positive values.