[version 1; peer review: 2 approved with reservations]
No competing interests were disclosed.
Combined hepatocellular cholangiocarcinoma (cHCC-CC) is a rare, but severely aggressive primary liver cancer manifesting characteristics of both hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC). The incidence rate is approximately 0.59 per 1,000,000 populations worldwide
CC is generally asymptomatic in early stages and is usually diagnosed late when the disease has already metastasized. Late-stage diagnosis limits the effective therapeutic options and has an aggressive disease course
At present, medical cannabis products are in use in many countries.
This study was reviewed and approved by the Mahasarakham University Human Research Ethics Committee (Reference No. 204/2563; approved on July 24, 2020), and Roi-Et Regional Hospital (Reference RE064/2563; approved on August 26, 2020), Buriram Regional Hospital Ethics Committee for Human Research, based on the Declaration of Helsinki and the ICH Good Clinical Practice Guidelines (Reference No. GCP0066/2563; approved on February 4, 2020). Because of its retrospective manner, informed consent was waived by the Roi-Et Regional Hospital and Buriram Regional Hospital. Data were collected from August 30, 2020, to June 30, 2021, which collected event data on 491 cases from September 1, 2019, to June 30, 2021.
A retrospective cohort study was conducted with 491 cHCC-CC patients (404 received ST and 87 received CT), who were diagnosed at least by ultrasonography and treated by supportive treatment at a palliative care and/or cannabis care clinic between September 1, 2019, and June 30, 2021. Data were extracted from the four tertiary hospitals, and two secondary hospitals serving five provinces of northeastern Thailand (Roi-Et Regional Hospital, Buriram Regional Hospital, Surin Provincial Hospital, Sawang Dandin Crown Prince Hospital, Panna Nikhom Hospital and Pana Hospital).
The independent variables were age at registration (palliative clinic and/or cannabis clinic), sex, cancer treatment and period of diagnosis to registration. The dependent variable was the survival time of patients with cHCC-CC. In order to calculate survival time, the starting point was identified as the date of registration, and the follow-up period ended when a patient passed away or on completion of the study. Censored data were used for those still alive at the end of the study (June 30, 2021) or lost to follow-up. Death status on the cause of treatment was confirmed by linkage with the death certificates from the national statistics database and by a telephone call to the patient or public health officer in community health centers.
Statistical analysis was performed with
ST, standard palliative care pain management treatment group; CT, medicinal cannabis treatment group.
| Variable | Patient treatment group | Median time, month (95% CI) | Person-time, month | Incidence rate/100 person/month | HR adj. 95% CI | P-value | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ST (n=404, %) | CT (n=87, %) | ST (n=404, %) | CT (n=87, %) | P-value | ST | CT | ST | CT | |||
|
|
0.83 (0.71–0.95) | 5.66 (1.94–9.38) | <0.001 | ||||||||
|
|
66.60 (11.67) | 65.64 (9.82) | <0.001 | ||||||||
| <60 | 105 (25.99) | 24 (27.59) | 0.83 (0.60–1.00) | 5.67 (2.87–15.00) | 170 | 147 | 0.59 | 0.08 | 1 | 0.212 | |
| 60–69 | 121 (29.95) | 28 (32.18) | 0.93 (0.73–1.04) | 3.27 (2.0–12.00) | 244 | 128 | 0.47 | 0.13 | 0.85 (0.66–1.09) | ||
| ≥70 | 178 (44.06) | 35 (40.23) | 0.83 (0.67–1.27) | 6.00 (2.33–10.03) | 375 | 200 | 0.44 | 0.11 | 0.87 (0.68–1.09) | ||
|
|
|||||||||||
| Male | 242 (59.99) | 54 (62.07) | 0.73 (0.67–0.93) | 6.00 (3.07–10.03) | <0.001 | 427 | 300 | 0.53 | 0.10 | 1 | 0.236 |
| Female | 162 (40.01) | 33 (37.93) | 0.97 (0.83–1.20) | 3.50 (1.77–9.50) | 362 | 175 | 0.42 | 0.11 | 0.89 (0.73–1.08) | ||
|
|
|||||||||||
| Surgery | 28 (6.93) | 4 (4.59) | 1.33 (0.30–2.50) | 2.00 (1.83–10.00) | <0.001 | 106 | 14 | 0.20 | 0.21 | 1 | 0.106 |
| Chemotherapy | 140 (34.65) | 18 (20.70) | 0.93 (0.73–1.0) | 9.50 (5.17–15.00) | 209 | 139 | 0.65 | 0.06 | 1.43 (0.93–2.2) | ||
| Combine | 149 (36.88) | 22 (25.29) | 0.83 (0.67–1.27) | 7.00 (1.67–15.00) | 311 | 121 | 0.45 | 0.09 | 1.27 (0.82–1.93) | ||
| Palliative care | 87 (21.54) | 43 (49.42) | 0.73 (0.5–0.93) | 3.07 (2.17–.8.33) | 162 | 201 | 0.51 | 0.14 | 1.23 (0.79–1.92) | ||
|
|
|||||||||||
| ST | 404 | 87 | 0.83 (0.71–0.95) | <0.001 | 1 | <0.001 | |||||
| CT | (82.3) | (17.7) | 5.66 (1.94–9.38) | 0.28 (0.20–0.37) | |||||||
|
|
|||||||||||
| Mean (SD) | 6.12 (2.55) | 5.46 (2.94) | <0.001 | ||||||||
| < 3 months | 60 (85.14) | 40 (45.98) | 0.93 (0.67–2.00) | 3.17 (2.17–9.00) | 115 | 115 | 0.54 | 0.14 | 1 | 0.844 | |
| 3–6 months | 204 (49.50) | 22 (25.28) | 0.83 (0.67–0.97) | 8.17 (2.87–15.00) | 406 | 406 | 0.46 | 0.08 | 1.31 (1.01–1.71) | ||
| 6–9 months | 94 (27.23) | 8 (9.20) | 1.07 (0.67–1.67) | 5.00 (0.73–8.00) | 210 | 210 | 0.41 | 0.09 | 1.21 (0.89–1.65) | ||
| >9 months | 46 (39.11) | 17 (19.54) | 0.67 (0.44–1.77) | 5.17 (200–9.00) | 59 | 59 | 0.72 | 0.09 | 1.16 (0.82–1.63) | ||
|
|
|||||||||||
| Passed away | 382 (94.60) | 52 (59.80) | |||||||||
The survival rate data after registration at either the palliative or cannabis care clinic. The cumulative 3, 6, 9 and 12 months survival rates were 28.80% (95% CI: 24.72–32.99), 20.00% (95% CI: 16.35–23.92), 16.50% (95% CI: 12.86–20.55) and 15.75% (95% CI: 12.04–19.92) for ST, 60.48% (95% CI: 49.35–69.91), 48.63% (95% CI: 36.78–57.70), 35.73% (95% CI: 23.83–47.74) and 29.98% (95% CI: 18.15–42.73) for CT, respectively. The median duration of survival was 0.83 months (95% CI: 0.71–0.95) for ST and 5.66 months (95% CI: 1.94–9.38) for CT. None of the demographic factors were significantly associated with survival time for either ST or CT. Comparing ST with CT, there was a statistically significant difference in age, sex, cancer treatment and period diagnosis with advanced cHCC-CC to register factors (p-value<0.05). There were factors found that affected the survival of patients receiving palliative care for liver and bile duct cancer. The most significant treatment factor found was between those patients who received standard therapy and those who received medical cannabis. Those on standard therapy were 3.57% more at risk of death than those on cannabis.
Multivariate analysis showed that CT treatment protocol was associated with improved patient survival, which was statistically significant (P value <0.001, the median time of CT, 5.66 months (95% CI: 1.94–9.38) and ST, 0.83 months (95% CI: 0.71–0.95), HR adj, 0.28 (95% CI: 0.20–0.37).
The impact of two types of treatment that affect the survival of cHCC-CC patients who either had supportive treatment at palliative clinic or a cannabis clinic. CT was the most effective treatment, with an overall survival time of 5.66 months, while overall survival time was 0.83 months for ST. Meanwhile, the overall survival times are consistent with other findings for after supportive treatment
In the univariate analysis, cancer treatment and period of diagnosis with advanced cHCC-CC to registration were associated with survival rate. It was found that the ST registered patients survived less than three months after being diagnosed with advanced-stage cHCC-CC. The reason for this might be that some patients had been consulting and were being cared for by an oncologist or other doctor rather than those patients who were registered for and receiving supportive treatment at a Palliative Clinic. Furthermore, most patients had been treated with a combination of surgery and chemotherapy, before being admitted to a Palliative Clinic. Although the registered patients at the Cannabis Clinic were >70 years old, they had no cancer treatment, only supportive treatment at the Cannabis Clinic. At the community hospitals where CT/MRI/biopsy/US have shown advanced organ metastases others who received treatment at a Cannabis clinic without waiting for a consultation with an oncologist were able to receive chemotherapy along with cannabis. This study has several limitations. A number of patients in the CT group dropped out before completion of the study. As a consequence, most patients suffering from advanced cancers and receiving heavy oncological treatments were older adults.
Patients with CCA have poor prognosis and short-term survival at the time of diagnosis. Registration and decision-making at the standard and/or cannabis clinic in each hospital differs across physicians, patients, families, stages of disease, organ metastasis, methods of treatment, and severity of symptoms. To the best of our knowledge, this is the first study that has compared survival rate and quality of life of CHCA/CCA patients who received either ST or CT across tertiary and secondary hospitals and across five provinces. Medical cannabis used in this study were standardized cannabis preparations made by the Thailand Food and Drug Administration. The side effects, safety, benefits and harms of the cannabis produced have been reviewed and are considered appropriate patient treatment. Prescribing doctors are trained, registered prescribers of medical cannabis.
N.P., contributed to Conceptualization, Data Curation, Formal Analysis, Resources, Methodology, Investigation, Writing – Original Draft. P.P., and A.W., contributed to Methodology, Investigation, Resources, Validation, Formal Analysis, Visualization. N.T., contributed to Conceptualization, Investigation, Supervision, Visualization, Writing – Review & Editing. M.A., contributed to Investigation, Visualization, Writing – Review & Editing. R.W., contributed to Conceptualization, Project Administration, Methodology, Investigation, Writing – Review & Editing, Funding Acquisition, and Supervision.
Figshare: Data_survival_cannabis.
Figshare: F1000_survival_table1_narisara_ranee.
Data are available under the terms of the
We would like to express the appreciation for all patients who participated in this research and the Hospital Center of Excellence Team (palliative clinic and cannabis clinic) for their invaluable help and encouragement throughout the course of this research. This research project was financially supported by Mahasarakham University. An earlier version of this article can be found on Research Square (doi:
Overall, this piece of work on the benefits of medical cannabis in improving the survival rate of combined hepatocellular cholangiocarcinoma (cHCC-CC) patients is interesting. However, several issues need further clarification and improvements.
It is important to recheck the accuracy of the provided data (
The authors did not mention whether they performed a proportional hazards assumption test in the statistical analysis. Upon re-analysis, I found that the treatment variable violated this assumption, indicating that the hazard ratio was not constant at 0.28 across the entire follow-up time as reported. More importantly, the significant association between medical cannabis and survival rate was observed only in the early follow-up (3 months), not the whole study period. Therefore, it is important for the authors to conduct a re-analysis and introduce an interaction term between the treatment variable and time using the time-varying covariate (TVC) option in STATA. When reporting the hazard ratio, the authors should present it across the range of follow-up (e.g., within 10 months) to ensure validity and reliability. More details and examples can be found at Bellera
Please avoid making causal claims in an observational study. For example, the sentence "the medical cannabis increased overall survival rates among cHCC-CC patients" (conclusion of the abstract) should be revised to "the medical cannabis was associated with improved overall survival rates among cHCC-CC patients." This distinction is vital because several alternative explanations (e.g., bias, errors, confounding, effect modification, reverse causality) could account for the significant findings. For instance, patients who received medical cannabis might have been more closely monitored by physicians, or they might have had unobserved or unmeasured characteristics (residual confounders) that affected their prognosis positively. Additionally, there may have been discrepancies in the quality of care across different settings. Therefore, the authors should refrain from assuming causality to avoid exaggerating the significance of their results. It would be appropriate to include a cautionary statement in the discussion section, such as "as this is an observational study, we cannot infer causality, and a randomised controlled trial is needed to establish the efficacy of medical cannabis in cancer patients." Please provide justification and references for each covariate selected as adjusting factors in the analysis. Furthermore, clarify why certain continuous variables (e.g., age, disease duration after registration) were categorised instead of using them as continuous scales. Additionally, I recommend running the analysis with age as a quadratic term (age + age
2), as it was found to be significantly associated with death, suggesting a non-linear relationship between age and mortality. Therefore, using a quadratic term for each continuous variable would be more appropriate as it preserves important information
Provide more details about the medical cannabis used in the study, such as product details, dosage form, dose, and administration. This information is crucial for generalising the findings to a clinical setting and enabling reproducibility. Additionally, clarify what the standard treatment was in the study and whether it was consistent across different settings. In the discussion section, provide more information about the individuals who were lost to follow-up and discuss how their exclusion might have influenced the findings. Is it possible to determine whether these patients were still alive at the end of the study or if they died soon after dropping out? Discuss the potential impact of differences in the quality of care across settings on the survival rate of patients in the study. It would be helpful to perform a subgroup analysis according to settings and utilise the strata option in the Cox model. Furthermore, consider conducting subgroup analyses based on other variables such as sex, age group, and current treatment to assess whether effect modification plays a role in the findings. Sensitivity and subgroup analyses are necessary to ensure the robustness of the findings, particularly in an observational study. In the discussion section, compare the survival rates of the study, particularly in the standard treatment group, with previous works. If applicable, discuss the reasons for any differences observed. This will help strengthen the external validity of the study. Use "multivariable" instead of "multivariate" when discussing regression models. The term "multivariable" refers to adding explanatory variables (X) in the regression model, while "multivariate" implies examining various outcomes (Y) simultaneously
Spell "proportional hazards regression" with an "s" in "hazards" since the term "proportional" implies the existence of at least two hazards. Be aware of the term person-months as it is not a person per month, but it is rather the product of patients and their corresponding follow-up time. So, the unit of the incidence rate in your work should be written as “100 person-months” not “100 person/month”. To improve clarity, consider splitting Table 1 into three separate tables. Table 1 should focus solely on the characteristics of included participants, allowing for inferential statistics (e.g., independent t-test, chi-squared test) to test the association between each characteristic and exposure status. Then, create Table 2 to present details of the outcome variable according to exposure status. Finally, present Table 3 as the main findings regarding the association between treatment and all-cause mortality, including both crude (unadjusted) and adjusted hazard ratios. Additionally, including a Kaplan-Meier plot with a risk table would aid in visualising the survival rates between patients receiving medical cannabis and those receiving standard treatment. If possible, please discuss the potential biological mechanisms or underlying explanations of how medical cannabis can improve the survival rate of cHCC-CC patients.
Overall, addressing these major and minor points will greatly enhance the clarity, validity, and reproducibility of your study.
Is the work clearly and accurately presented and does it cite the current literature?
Partly
If applicable, is the statistical analysis and its interpretation appropriate?
Partly
Are all the source data underlying the results available to ensure full reproducibility?
Partly
Is the study design appropriate and is the work technically sound?
Yes
Are the conclusions drawn adequately supported by the results?
Partly
Are sufficient details of methods and analysis provided to allow replication by others?
Partly
Reviewer Expertise:
Clinical epidemiology, pharmacoepidemiology, and social epidemiology
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Based on the STROBE criteria, most of this research has fulfilled the criteria. In my opinion, there are several things that need to be improved so that this research is better:
There is no sufficient detail of the methods, especially on the eligibility criteria of participants, and the method of follow up that provided to allow replication by others. There are no very important data as a confounder, namely the pain scale/level of pain on the criteria when matching participants. It would be clearer if the authors can explain the relationship between pain levels, quality of life, and survival rates in biomolecular terms. Such as, for example, mutations in the NMDA receptor in chronic pain, which will produce P protein which can increase the risk of advanced metastasis, NMDA receptors stimulates the MAPK and CaMK pathways, leading to CREB activation in tumor cells. NMDAR-interacting proteins and the downstream signaling effectors display features in common between the neuronal and metastatic cancer processes, such as cell adhesion, migration, and survival. In the results of the cohort study, it would be better if the authors can explain in the report the numbers of individuals at each stage of study – e.g., numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and the reason of dropped out participant. Give reasons for nonparticipation/dropped out participant in each stage. Consider use of a flow diagram.
Thank you.
Is the work clearly and accurately presented and does it cite the current literature?
Yes
If applicable, is the statistical analysis and its interpretation appropriate?
Yes
Are all the source data underlying the results available to ensure full reproducibility?
Partly
Is the study design appropriate and is the work technically sound?
Yes
Are the conclusions drawn adequately supported by the results?
Yes
Are sufficient details of methods and analysis provided to allow replication by others?
Partly
Reviewer Expertise:
Surgical Oncology, biomolecular, Immunology, epidemiology
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
There is no sufficient detail of the methods, especially on the eligibility criteria of participants, and the method of follow up that provided to allow replication by others.
There are no very important data as a confounder, namely the pain scale/level of pain on the criteria when matching participants.
It would be clearer if the authors can explain the relationship between pain levels, quality of life, and survival rates in biomolecular terms. Such as, for example, mutations in the NMDA receptor in chronic pain, which will produce P protein which can increase the risk of advanced metastasis, NMDA receptors stimulates the MAPK and CaMK pathways, leading to CREB activation in tumor cells. NMDAR-interacting proteins and the downstream signaling effectors display features in common between the neuronal and metastatic cancer processes, such as cell adhesion, migration, and survival.
In the results of the cohort study, it would be better if the authors can explain in the report the numbers of individuals at each stage of study – e.g., numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and the reason of dropped out participant.
Give reasons for nonparticipation/dropped out participant in each stage. Consider use of a flow diagram.