Adherence, Safety, and Effectiveness of Medical Cannabis and Epidemiological Characteristics of the Patient Population: A Prospective Study

Outcome Measures

Statistical Analysis

Continuous variables with normal distribution were presented as means with standard deviation. Ordinary variables or continuous variables with non-normal distribution were presented as medians with an interquartile range (IQR). Categorical variables were presented as counts and percent of the total.

We used a t-test for the analysis of continuous variables with normal distribution, Mann–Whitney U-test whenever parametric assumptions could not be satisfied, and χ²-test for categorical variables. For paired comparisons, we have used paired t-test, non-parametric Wilcoxon signed rank test, and χ²-test for dependent variables. We used multivariable logistic regression analysis for the factors associated with the treatment success to control possible confounders. We have included the following baseline variables into the models based on clinical considerations: age, gender, weight, indication for cannabis treatment, presence of pain, number of chronic medications, hospitalization in the past 6 months, employment, car use, previous experience with cannabis, cigarette smoking, QOL, and concerns about cannabis treatment as reflected in the intake form. The final model was selected according to the model characteristic, evaluated by calculating the c-statistic, in addition to choosing the minimal −2 log likelihood of each model.

Results are displayed as odds ratios with 95% confidence interval. P-value < 0.05 was considered statistically significant. All analyses were performed at the Clinical Research Center, Soroka University Medical Center, Beer-Sheva, Israel, using IBM SPSS version 22 (SPSS, Chicago, IL). The Sponsor was not involved in the data analysis.

Adherence

Adherence was assessed for all patients who initiated treatment in the cannabis clinic. After1 month, of the 9,985 patients who started the treatment, 4.8% died, 5.2% stopped treatment, and 0.3% switched to a different cannabis supplier, while 89.7% continued active treatment. Of those who continued active treatment, 6,699 (74.8%) responded to the questionnaire. Of them, 2,562 patients (38.2%) experienced side-effects or reported that the cannabis did not improve their condition during the first month of the treatment and needed the advice and guidance of a nurse to adjust the dose or the treatment. At 6 months of 7,773 patients, 6,038 (77.7%) remained in active treatment (excluding 19.4% patients who died and 2.7% who switched to a different cannabis supplier).

Safety Analysis

Of the 4,891 patients who responded to the side-effect follow-up questionnaire, 1,675 patients (34.2%) reported experiencing at least one side-effect. The most common were dizziness (8.2%), dry mouth (6.7%), increased appetite (4.7%), sleepiness (4.4%), and psychoactive effects (4.3%) (Table 3). This analysis included all active patients and patients who discontinued the cannabis treatment.

Increased appetite was reported as a side effect by 232 patients (4.7% overall and 2.0% as a lone side effect); 36.6% of them received their cannabis license for pain indication, 34.4% for cancer, 11.7% for PTSD, 4.3% for Crohn’s and colitis, and 3.9% for autism.

Of those responding to the side-effects chapter, 2.9% reported nausea. This rate varied between different chemovars in the interval of 1.2–3.8%, with THC rich indica chemovar “Dorit” being the highest.

Primary Effectiveness Outcome

The primary effectiveness outcome was assessed for all respondents to the intake questionnaires except for patients refusing treatment (69), deceased patients (1,785), patients switching to other providers (271), and active patients who did not respond to the follow-up questionnaire (911). Thus, the primary effectiveness outcome was assessed for 5,524 of the 8,560 patients responding to the intake questionnaire (64.5%, Figure 1). Overall, 3,902 (70.6%) patients out of 5,524 experienced treatment success. Multivariable analysis revealed the following factors as associated with treatment success: cigarette smoking (O.R 2.4, 95% C.I 2.0–2.2), prior experience with cannabis (O.R 2.1, 95% C.I 1.8–2.5), driving (O.R 1.3, 95% C.I 1.1–1.5), employment (O.R 1.3, 95% C.I 1.0–1.5), and young age (O.R 0.9, 95% C.I 0.9–0.9), (Table 4).

Of 4,364 patients who answered to the 6-months follow-up questionnaire, the most common chemovar used was an 18% THC indica (Erez, 2,551 patients, 55.7%). This chemovar was most often consumed by smoking or vaporization (1,306 patients), at an average dried flowers weight of 0.3 g (54 mg THC) per administration, and a frequency of 3.4 administrations per day. A total of 935 patients consumed Erez sublingual oil 300 mg THC/10 ml, consuming an average dose of 5.7 mg THC per administration, and a frequency of 2.4 administrations per day (further description in Table 5).

The most improved symptoms were rage attacks (decrease of 91.5%), restlessness (89.5%), sleep disturbances (89.1%), and nausea (88.9%). For more information about the changes in the specific symptoms after 6 months, please refer to Supplementary Table 4. Of the patients reporting nausea at intake and responding to the follow-up questionnaire, 30.5% reported that they no longer suffer from nausea, 58.4% reported that the symptom improved, 10.0% reported no change in nausea they experience, and 1.1% of patients reported deterioration in the nausea they experienced.

Pain intensity was assessed both at intake and at 6 months in 4,166 patients. Prior to treatment initiation, 62.0% of patients reported their pain at between 8 and 10, while only 5.0% reported this intensity at 6 months (p < 0.001, Figure 2); 7.3% of the patients demonstrated deterioration in their pain scale. In 17.8%, the level of pain did not change while in 74.7% it improved, of which 64.3% of patients showed an improvement of 30% or more in their reported pain intensity and 47.2% reported an improvement of 50% or more in their pain intensity. In 1,580 patients, only under the pain indication, 85.9% experienced an improvement of 30% or more, and 59.3% an improvement of 50% or more in their VAS pain scale.

Pain intensity was assessed both at intake and at 6 months in 4,166 patients. Prior to treatment initiation, 62.0% of patients reported their pain at between 8 and 10, while only 5.0% reported this intensity at 6 months (p < 0.001, Figure 2); 7.3% of the patients demonstrated deterioration in their pain scale. In 17.8%, the level of pain did not change while in 74.7% it improved, of which 64.3% of patients showed an improvement of 30% or more in their reported pain intensity and 47.2% reported an improvement of 50% or more in their pain intensity. In 1,580 patients, only under the pain indication, 85.9% experienced an improvement of 30% or more, and 59.3% an improvement of 50% or more in their VAS pain scale.

Quality of life (QOL) was assessed both at intake and at 6 months in 4,143 patients. While only 12.9% of patients reported good QOL prior to treatment initiation, 69.9% reported good QOL at 6 months (p < 0.001, Figure 3).

Figure 3

Quality of life assessment. Quality of life was assessed prior to and 6 months after initiation of cannabis treatment (p < 0.001). Assessment was done on 4,143 patients who responded twice to this question.

Concomitant medications consumption was evaluated both at the intake and in the follow-up questionnaires in 3,544 patients. The most reduced medications classes were opioids (52.5%), other analgesics and antipyretics (39.2%), anti-psychotics (36.9%) anti-epileptics (35.7%), and hypnotics and sedatives (35.3%) (Table 6).

Figure 4A presents rates of the primary outcome of treatment effectiveness and safety at 6 months, stratified by the primary indication for use, ranging between 55.4% for epilepsy to 90.8% for PTSD. Figure 4B presents the proportion of patients experiencing any side-effect, and ranges between 28.9% for Tourette syndrome to 40.0% in patients with epilepsy.

In the analysis of the subgroup of 1,500 patients with only one chemovar used, we found significant differences in success rate. The two chemovars utilized by more than 50 patients that differ in the success rate were Alaska 91% chance of success vs. Avidekel with 66.4% (p < 0.001). However, these chemovars were utilized by patients with different medical conditions and therefore the direct comparison of the success rates is not fully informative. We have not found a difference of safety rates between the different chemovars (p > 0.05).

Demographic Characteristics

The characteristics of medical cannabis users (age, severity of diseases, number of comorbidities, number of symptoms, number of medications, employment status, etc.) depend upon and are determined by the limitations and laws set by governmental and regulatory authorities. During the study period, Israeli national regulation of medical cannabis provided strict guidelines for the patients and their physicians on the use of the medication. The demographic and medical characteristics of our cohort differ from most reported populations. The Israeli medical cannabis patients are on average (55 years old) two and a half-decade older than patients in comparable reports (2, 8, 10, 12–16, 18–21), with a more balanced gender distribution (51.3% men compared to 60–80% in most studies) (2, 11, 13–19, 21–24). In the current cohort, the main indication for cannabis treatment was cancer (48.9%), while in other studies the main indications were pain (2, 10–12, 15, 18, 24, 25, 33–35), anxiety (13, 14, 36), and depression (19); cancer was diagnosed in only 7.4–11.4% of the patients (2, 10–12, 14, 15, 19, 24, 25).

Almost 20% of the study population died within the first 6 months of follow-up primarily due to malignancies (90.1%).

Treatment Adherence

Adherence to cannabis treatment was 77.7%, similar to the treatment withdrawal of 23.8% that was found in a retrospective cohort study on medical cannabis patients with a mean age similar to the patients’ ages in our study (33).

Treatment adherence in our cohort was favorably comparable to the expected adherence in patients taking chronic medications: in a systematic review of 76 studies, patients taking medication on a schedule similar to the cannabis treatment regimen of at least four times daily, demonstrated average adherence rates of 50% (range 31–71%) (37). Furthermore, in a study of long-term treatment with opioids, treatment was discontinued in 51% of the patients (38).

Safety

The safety of cannabis treatment in this heterogeneous population of patients was found to be high, especially when compared with the safety of long-term opioid treatment. Side-effects of medical cannabis were infrequent, minor, and rarely the cause of discontinuation. The most common side-effect, dizziness, was reported by 8.2% of the active responders, while the usual prevalence of side-effects in patients on opioid therapy is substantially higher: more than 40% of the patients report dizziness, more than 35% report constipation, more than 30% report nausea, and more than 25% report fatigue (38). In addition, long-term opioid treatment is associated with sedation, cognitive impairment, depression, addiction (39), and subtle neuropsychological changes (40–42). These high-safety results are similar to a large, controlled study that prospectively measured the safety of a high-THC medical cannabis product in 215 patients treated in chronic pain clinics. The patients were compared with 216 patients in the clinics who did not use medical cannabis and were followed-up for 1 year. The adverse events in this study were modest, and no significant difference in the occurrence of serious adverse effects was found (43). These results may be attributed to the safety-focused approach implemented; a guided choice of chemovar and route of administration, a slow titration method, an initial follow-up after 1 month, and a follow-up after 6 months, could be the strategy that ensured that harms from medical cannabis were mitigated (44).

Effectiveness

Although this study is observational and thus no causality can be established, the treatment seems effective in reducing pain, in increasing QOL, and in reducing concomitant medication consumption. In our cohort, the primary effectiveness outcome was achieved by more than 70% of the patients, while only 17.4% of the patients discontinued treatment. Although further head-to-head comparative study between opioids and cannabis for palliation is needed, our results demonstrate numerically comparable effectiveness in pain treatment (e.g., opioids treatment provides adequate relief for 70–90% of patients with cancer pain) (45). However, long-term opioid treatment in non-specific pain patients delivered good pain relief in only 51% of patients (37). Although Cochrane review of neuropathic pain treated with cannabis-based medicines against placebo, found a modest gain from 16 studies (n = 1,750) with 21 vs. 17% achieving a 50% reduction in pain; and 39 vs. 33% achieving a 30% reduction (46), a multiple-criteria decision analysis found that the benefit-safety profiles for cannabinoids were higher than for other commonly used medications for chronic neuropathic pain largely because they contribute more to QOL and have a more favorable side-effect profile (47). Furthermore, for patients with chronic pain, opioids may contribute to substantial functional impairment (48), so serious adverse effects of opioids may limit effectiveness in some patients, even if adequate analgesia is achieved (48). The lack of serious side-effects of broad-spectrum cannabis products together with the effectiveness albeit shown in the small studies makes cannabis products a possible alternative for the treatment chronic pain.

The fact that previous experience with cannabis was associated with a higher chance of treatment success, can suggest that the placebo effect contributed to the overall improvement, as an expectation of a positive influence may increase the magnitude of the placebo effect. Moreover, young patients (usually with fewer comorbidities) that drive, smoke cigarettes, and are employed seems more likely to experience and report improvement following treatment. It is also possible that patients who smoke cigarettes know how to perform the inhaling action and are more likely to benefit from the treatment.

The broad effect of medical cannabis treatment, which has a beneficial effect on a variety of symptoms, can potentially explain the reduction in drug consumption (especially of painkillers). Cannabis may be a viable alternative to opioids for those experiencing pain (49, 50).

Out of 1,160 patients responded to the intake questionnaire and discontinued treatment, 593 filled the follow-up questionnaire at 6 months. The most common reasons for discontinuing treatment were side-effects (25.0%), no therapeutic effect (24.6%), no longer a need for cannabis treatment (23.2%), or failed renewals of mandatory cannabis treatment licenses (6.8%). Furthermore, 44.3% of the patients who discontinued the treatment have reported at least moderate improvement in their symptoms following cannabis treatment. Even though all patients who discontinued treatment were classed as “treatment failures”, we have recorded high rates of treatment success.

Treatment with medical cannabis is complex for several reasons: (1) the multiplicity of potential treatment chemovars, (2) the multiplicity of consumption options, (3) and because each patient will receive a different therapeutic dose, patients need to “find” their therapeutic dose in a slow titration process that is dictated by the psychoactive effect and other treatment side-effects.

A significant percentage of patients expressed concerns about initiating cannabis treatment. In addition, in the short-term follow-up (after about a month of active treatment), a large group of patients needed additional consultation with an experienced cannabis clinic nurse in order to adjust the dosage or the treatment product, emphasizing the great importance of professional guidance and instruction during the first ~2 months of treatment. Without guidance, patients may take too high a dose, experience a side-effect, and abandon the treatment. In addition, without setting expectations regarding the patience required in the first weeks of treatment (until the body adapts to the product, and until reaching the therapeutic dose, especially with CBD products), the patient may conclude that, if after several attempts his condition does not improve, the treatment is unhelpful and so may eventually quit.

We would like to acknowledge and thank the participants and their families for their participation and support.