I have used the following eligibility criteria: (I) studies should report the characteristics of patients diagnosed with opium dependence; (II) studies should report characteristics of patients diagnosed with any kind of stroke; (III) studies should report the association between opium with stroke incidence, prevalence, and outcome; and (IV) studies should be observational (namely, cohort, case-control, or cross-sectional studies).

A systematic search was conducted through PubMed, Scopus, and Web of Science (WoS), from inception, until January 12, 2022. I also carried out a manual search using google scholar.

The search strategy for the systematic search was designed comprising two concepts; stroke and opium. I searched (“stroke” AND “opium”) OR (“cerebrovascular” AND “opium”) OR (“cva” AND “opium”) OR (“apoplexy” AND “opium”) OR (“stroke” AND “papaver”) OR (“cerebrovascular” AND “papaver”) OR (“cva” AND “papaver”) OR (“apoplexy” AND “papaver”). The search strategy algorithm is also shown in Supplementary Table 1.

The study selection was carried out via the EndNote reference management software in three phases to manage identified records. Initially, duplicate articles were removed using the endnote software and manually, followed by the screening phase, in which the title and abstract of the studies were evaluated based on the inclusion criteria. Afterward, the screening of the full texts was conducted.

I filled data extraction sheets containing variables such as author name, year of publication, age, gender, study type, sample size, the definition of exposure and outcome, the measure of association, and study findings.

Quality assessment (QA) of included studies was conducted by STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guideline, comprising 22 sub-items. Each item's ratings are yes (1) or no (0), and the total QA score is the sum of these sub-items. This guideline restricts its recommendations to the three main analytical designs used in observational research (cohort, case-control, and cross-sectional studies) (16).

The findings of this study are presented as odds ratios (ORs) and 95% confidence interval (95% CI). In order to perform the meta-analysis, STATA version 14.0 (StataCorp, College Station, TX) software was used. I restrict the meta-analysis when two or more records report the association of opium and stroke. The pooled estimate of ORs and their 95% CI was calculated based on the extracted datasheet. I² statistic and chi-square-based Q-test were used for the assessment of heterogeneity. In the current study, the lack of heterogeneity was specified when the p-value was more than 0.10. Random models were used to pool the ORs. Publication bias was measured using Begg's test. Publication bias was considered substantial whenever the p-value was <0.1.

Our manual and systematic searches yielded 110 studies from PubMed, 88 studies from Scopus, 92 studies from Web of Science, and 302 from Google scholar. After duplicates removal, I evaluated 584 studies for eligibility, followed by the exclusion of 566 records based on their title and abstract. Finally, eight and four studies were included in the qualitative and quantitative syntheses. The detailed flow diagram is shown in Figure 1.

Table 1 demonstrates the characteristics of included studies (3, 17–23). Six (75%) studies included in the qualitative analysis were case-control, while the remaining two records were cohort studies. The sample size of included studies ranges from 133 in the Mousavi-Mirzaei et al. study (21) to 50,045 in the Khademi et al. study. The highest percentage of men in a paper was 57.8 in Saberi et al. study. QA of Studies based on STROBE showed that the QA score of included studies ranges from 15 to 20 out of 22.

This study indicated that not only opium dependence almost doubles the odds of being diagnosed with a stroke but also it leads to a higher hazard of mortality due to stroke. This is the first systematic review and meta-analysis that proposes opium dependence as a notable risk factor for stroke to the best of my knowledge. It should be clarified that this study and included records are all focused on the cerebrovascular effects of opium abuse, rather than the effects of prescribed opium, as a component of a medical treatment plan.

Our meta-analysis on multi-variable adjusted case-control studies revealed that opium consumption independently increases the odds of ischemic stroke. Moadabi et al. proposed hypothetical pathophysiology underlying the impact of opium on cerebral hemodynamic parameters. They performed transcranial Doppler (TCD) ultrasonography, which revealed that patients with opium dependence are at higher odds of abnormal pulsatility index (OR = 9.615) and mean flow velocity (OR = 3.246). In other words, they indicated that opium affects both small and large vessels, which leads to a higher risk of ischemic stroke (24). Likewise, Mousavi-Mirzaei et al.'s study revealed that opium consumption leads to increased atherosclerotic plaques and internal media thickness, a well-established component of the pathogenesis of ischemic stroke (21). These findings are similar to studies showing an increased risk of myocardial infarction and coronary revascularization, among opium users (25). Opium consumption is a significant risk factor for coronary artery disease, as it increases the progression of atherosclerosis, and leads to an increase in low-density lipoprotein cholesterol (LDL-C) level and a decrease in high-density lipoprotein cholesterol through opium-induced activation of hepatic pathways (26–28).

Moreover, the Asiabanha et al. study demonstrated that opium imposes brain cell damage by inducing apoptosis, which may prone the brain to cerebrovascular incidents (29). Opium addiction also affects endothelial function, leading to a decline in nitrous oxide level (30), a molecule that improves brain blood flow and neuronal survival in stroke (31).

Apart from previously mentioned pathways, opium addiction is a well-established risk factor for diabetes, hypertension, and hyperlipidemia (32), which can potentially elevate stroke hazards (33, 34). Studies, included in the current meta-analysis, were adjusted for potential confounders such as dyslipidemia, diabetes, and hypertension. Nevertheless, it should be noted that previous studies revealed that opium consumption worsens glucose metabolism in patients both with diabetes and non-diabetes (35, 36), and it results in a significant disturbance in lipid profile, which in turn may provoke stroke (28). To sum up, opium not only directly affects brain cells but also leads to unfavorable changes in the incidence of metabolic risk factors.

Another modifiable risk factor for stroke is tobacco smoking. Comparing the effects of tobacco smoking and opium as habitual risk factors can help understand the harmful but modifiable effects of opium. OR calculated for the effect of opium on stroke in this study was higher or similar compared to ORs calculated for tobacco smoking (37–41). Pan et al.'s meta-analysis revealed that current smokers are at higher odds of stroke (1.92, 95% CI: 1.49–2.48), while the association of being a former smoker and stroke was not significant (1.30, 95% CI: 0.93–1.81) (38). Likewise, Wannamethee et al. study showed that quitting smoking results in notable and rapid benefits in reducing the risk of stroke (42). Interestingly, smoking cessation even after stroke enhances the outcome (43). Beneficial effects of opium cessation are also evident in previous studies (3, 44), although there is no scientific evidence regarding the impact of opium cessation and a decline in stroke risk.

Smoking is generally more prevalent in men (45). Similarly, the prevalence of opium addiction is notably higher in men than women (3). As the number of studies included in my meta-analysis was few and no heterogeneity was evident, I could not conduct a meta-regression analysis to evaluate the effect of gender. However, it is worth noting that experimental studies showed that addicted men are at higher risk of ischemic stroke as men brain cells are more susceptible to opium than women (29).

Our review shows that opium predisposes patients to stroke and increases the hazard of mortality following any type of stroke. In other words, Khademi et al.'s study demonstrated that opium consumption leads to strongly increased risks of death due to a stroke. That is to say, in GCS, opium was the cause of mortality in 0.50% of non-users compared to 1.15% of participants who consume opium. These findings remained significant after adjustment for potential confounders (3).

Comparably, Andalibi et al.'s study illustrated that using opium increases mortality risk following stroke. More specifically, multi-regression models proved that patients who consume opium are at higher hazard of 3-months (HR = 1.58, 95% CI: 1.01–2.49), 1-year (HR = 1.78, 95% CI: 1.20–2.65), and 5-years (HR = 1.75, 95% CI: 1.28–2.40) mortality. Conversely, patients' disability was not significantly different in user and non-user groups. Participants' disability was measured using an objective index (Modified Rankin Scale), which can be influenced by the analgesic and euphoric effects of opium (17).

Given the remarkable effects of opium on stroke, addiction prevention programs should be seriously pursued in countries where opium dependence is common, namely, central Asian countries and Iran (46). Although I did not add a filter for the provenance of the study, all included studies were from Iran in the current review. Even though the single geographical provenance of included studies makes the generalization of my findings challenging, it can assist other low and middle-income countries, similar to Iran, in adopting better policies.

This study had several strengths. Most importantly, this is the first systematic review and meta-analysis that evaluates the effects of opium on stroke. This study evaluates the effect of opium on stroke incidence and mortality, and it also reviewed papers that evaluated the outcome of patients dependent on opium following stroke. This study also had several limitations; first, the number of studies included in the quantitative analysis was limited, resulting in a bias. Second, all included studies were from a single country (Iran). Third, studies included in the meta-analysis were all case-control studies. Conduction of larger cohorts is suggested to assess the effect of opium more precisely as a risk factor. Fourth, due to the absence of randomization, p-values (even underlying the confidence intervals) must be considered as a possible approximation to reality.