Self-Reported Cannabis Use and HIV Viral Control among Patients with HIV Engaged in Care: Results from a National Cohort Study
, Yu Li 6, Rachel Vickers-Smith 7, Stephen Crystal 8, Robert D. Kerns 9,10, Kirsha S. Gordon 10,11, Alexandria Macmadu 6, Melissa Skanderson 10, Kaku So-Armah 12, Minhee L. Sung 10,11, Fiona Bhondoekhan 6, Brandon D. L. Marshall 6 and E. Jennifer Edelman 11,13,*
Round 1
Reviewer 1 Report
In this study the authors examined the association between longitudinal measures of cannabis use and HIV viral load among PWH engaged in care in the Veterans Aging Cohort Study. The manuscript is well-written, and the tables are adequately presented. The analysis was done using appropriate statistical method. However, there are some concerns that I highly recommend to address:
- My first concern refers to the lack of information on the amount of cannabis consumed by the participants. What is considered minimum or maximum doses?
- I suggest adjusting the tittle to increase adequation with the showing results.
- I suggest rewriting the introduction in link with the results shown.
Author Response
Comment #1. My first concern refers to the lack of information on the amount of cannabis consumed by the participants. For example, what is considered minimum or maximum doses?
Response #1. In this study, the participants’ cannabis consumption was quantified by frequency rather than dose, as described in the Methods. Ultimately, this is a general limitation of cannabis research, emphasizing the need for a standardized “dose” or “unit” of cannabis consumption. Therefore, we noted this as a limitation and modified this to now read:
Page 11, paragraph 2: Third, we did not capture cannabis dosage, potency, consumption modalities, cannabis use disorder, or cannabis legalization status across states.
Comment #2. I suggest adjusting the title to show results.
Response #2. While we appreciate this suggestion and have considered it, we prefer to keep the original title, “Self-reported cannabis use and HIV viral control among patients with HIV engaged in care: Results from a national cohort study,” given our findings of a statistically significant association in unadjusted, but not adjusted analyses.
Comment #3. In addition, I suggest rewriting the introduction in the link with the results shown.
Response #3. Thank you for this suggestion; we believe the current introduction provides relevant background and rationale for the research question addressed in this manuscript. We would be happy to revise upon more specific guidance and at the Editor’s discretion.
Reviewer 2 Report
In this manuscript “Self-Reported Cannabis Use and HIV Viral Control among Patients with HIV Engaged in Care: Results from a National Cohort Study” by Bahji et al. authors used a large cohort of more than 2,500 PLWH followed for 16 years. This is a very interesting work that shows a lack of association between cannabis consumption and HIV load control. The manuscript is very well written, and I appreciated the limitations paragraph at the end. I have one comment I would like authors to respond.
- Did you attempt to do a first analysis without any categorization and adjustments? Pure correlative analysis between consumption and viral loads?
- Have you performed the analysis only including those who had undetectable baseline HIV load? It is possible this may be a confounding factor that is carried over the analysis.
Author Response
Comment #1. Did you attempt to do the first analysis without any categorization and adjustments? Pure correlative analysis between consumption and viral loads?
Response #1. Yes, the first analysis was a crude (unadjusted) logistic regression to assess the overall association between cannabis use and the presence of a detectable viral load. This is what is captured in the “unadjusted” models and results. In addition, we have added a figure to demonstrate these associations (see Reviewer #3, Comment #1).
Comment #2. Have you performed the analysis only including those who had undetectable baseline HIV load? This may be a confounding factor that is carried over the analysis.
Response #2. Our goal was to assess the association between cannabis use on HIV viral load among those engaged in care and receiving ART. Therefore, we restricted analyses to those with ART at baseline regardless of HIV viral load. In response to this question, we examined the baseline characteristics of those with an undetectable vs. detectable HIV viral load at baseline; we believe it would be overly restrictive to limit analyses to those with baseline undetectable viral load. However, we have provided an additional table (below) that compares the characteristics by baseline HIV viral load (<500 copies/mL vs. (≥500 copies/mL) status. We are happy to add these findings to the paper at the Editor’s discretion.
In addition, we have expanded our suggested directions for future research:
Page 11, paragraph 3: In addition, future research may consider whether cannabis use moderates opioid and other substance use among PWH engaged in care [32], explore the association between cannabis use and earlier stages of the HIV care continuum (e.g., ART receipt) consider more nuanced assessments of cannabis use, including dosage, modality, type, indication, and context (e.g., CUD, CD, or CA versus use alone).
Supplementary Table 3. Sociodemographic and clinical characteristics associated with cannabis use categories among people living with HIV receiving antiretroviral therapy in the VACS cohort (2002-2018) comparing those with an undetectable viral load (<500 copies/mL) to those with a detectable viral load (≥500 copies/mL).
|
Characteristic |
Overall |
<500 copies/mL (N=1690, 67.2%) |
>=500 copies/mL (N=825, 32.8%) |
Global |
|
Demographics |
|
|
|
|
|
Age, mean (SD) |
50.1 (8.8) |
51.1 (8.9) |
48.0 (8.3) |
<0.0001 |
|
Gender, n (%) |
|
|
|
0.936 |
|
Male |
2447 (97.3) |
1644 (97.3) |
803 (97.3) |
|
|
Female |
68 (2.7) |
46 (2.7) |
22 (2.7) |
|
|
Race/ethnicity, n (%) |
|
|
|
0.085 |
|
Non-Hispanic White |
524 (20.8) |
373 (22.1) |
151 (18.3) |
|
|
Black |
1646 (65.5) |
1078 (63.8) |
568 (68.9) |
|
|
Hispanic |
247 (9.8) |
171 (10.1) |
76 (9.2) |
|
|
Other |
98 (3.9) |
68 (4.0) |
30 (3.6) |
|
|
Education, n (%) |
|
|
|
0.420 |
|
High school or less |
987 (39.7) |
672 (40.2) |
315 (38.6) |
|
|
Some college or more |
1500 (60.3) |
998 (59.8) |
502 (61.4) |
|
|
Marital status, n (%) |
|
|
|
0.255 |
|
Never married |
610 (24.6) |
421 (25.3) |
189 (23.3) |
|
|
Married/living with a partner |
981 (39.6) |
666 (40.0) |
315 (38.8) |
|
|
Divorced/widowed |
885 (35.7) |
577 (34.7) |
308 (37.9) |
|
|
Housing instability ever, n (%) |
982 (39.3) |
632 (37.7) |
350 (42.5) |
0.019 |
|
Location of residence, n (%) |
|
|
|
0.357 |
|
Urban |
2368 (95.0) |
1589 (95.1) |
779 (94.7) |
|
|
Suburban |
70 (2.8) |
42 (2.5) |
28 (3.4) |
|
|
Rural |
56 (2.3) |
40 (2.4) |
16 (1.9) |
|
|
Annual income, n (%) |
|
|
|
0.475 |
|
<$11,999 |
1213 (49.8) |
803 (49.2) |
410 (51.0) |
|
|
$12,000-$49,999 |
1038 (42.6) |
698 (42.8) |
340 (42.3) |
|
|
≥$50,000 |
184 (7.6) |
130 (8.0) |
54 (6.7) |
|
|
Social Isolation Score |
|
|
|
0.150 |
|
<4 |
680 (27.0) |
472 (27.9) |
208 (25.2) |
|
|
≥4 |
1835 (73.0) |
1218 (72.1) |
617 (74.8) |
|
|
HIV-related factors |
|
|
|
|
|
CD4 cell count, cells/mm3, median (IQR) |
374 (228, 568) |
427.5 (276, 622) |
282 (144, 446) |
<0.0001 |
|
VACS Index 2.0 score, median (IQR) |
56 (46, 66) |
53 (44, 64) |
61 (53, 72) |
<0.0001 |
|
ART adherent, n (%) |
1551 (61.7) |
1158 (68.5) |
393 (47.6) |
<0.0001 |
|
Other health conditions and status, n (%) |
|
|
|
|
|
HCV co-infection |
929 (36.9) |
634 (37.5) |
295 (35.8) |
0.391 |
|
Any cancer |
524 (20.8) |
354 (21.0) |
170 (20.6) |
0.843 |
|
Anxiety symptoms |
905 (37.1) |
592 (36.1) |
313 (39.1) |
0.150 |
|
Depressive symptoms |
534 (21.5) |
334 (20.0) |
200 (24.4) |
0.012 |
|
Pain interference |
830 (33.3) |
542 (32.5) |
288 (35.0) |
0.199 |
|
Other substance use, n (%) |
|
|
|
|
|
Smokes cigarettes |
1915 (76.1) |
1305 (77.2) |
610 (73.9) |
0.070 |
|
Unhealthy alcohol use |
876 (34.8) |
604 (35.7) |
272 (33.0) |
0.171 |
|
Past year stimulants or cocaine |
538 (21.4) |
319 (18.9) |
219 (26.6) |
<0.0001 |
|
Prescribed opioid receipt, n (%) |
|
|
|
0.198 |
|
No opioid receipt |
1802 (71.7) |
1215 (71.9) |
587 (71.2) |
|
|
Short-term + low dose |
449 (17.9) |
303 (17.9) |
146 (17.7) |
|
|
Short-term + high dose |
40 (1.6) |
25 (1.5) |
15 (1.8) |
|
|
Long-term + low dose |
159 (6.3) |
97 (5.7) |
62 (7.5) |
|
|
Long-term + high dose |
65 (2.6) |
50 (3.0) |
15 (1.8) |
|
|
Prescribed benzodiazepine |
|
|
|
0.840 |
|
None |
2147 (85.4) |
1445 (85.5) |
702 (85.1) |
|
|
Low dose |
269 (10.7) |
177 (10.5) |
92 (11.2) |
|
|
High dose |
99 (3.9) |
68 (4.0) |
31 (3.8) |
|
|
Prescribed gabapentin |
|
|
|
0.462 |
|
None |
2240 (89.1) |
1501 (88.8) |
739 (89.6) |
|
|
Low dose |
127 (5.1) |
83 (4.9) |
44 (5.3) |
|
|
High dose |
148 (5.9) |
106 (6.3) |
42 (5.1) |
|
|
Prescribed antidepressant |
|
|
|
0.121 |
|
None |
1557 (61.9) |
1064 (63.0) |
493 (59.8) |
|
|
Short term |
314 (12.5) |
196 (11.6) |
118 (14.3) |
|
|
Long term |
644 (25.6) |
430 (25.4) |
214 (25.9) |
|
|
Site |
|
|
|
<0.0001 |
|
Atlanta |
403 (16.0) |
191 (11.3) |
212 (25.7) |
|
|
Bronx |
258 (10.3) |
181 (10.7) |
77 (9.3) |
|
|
Houston |
335 (13.3) |
212 (12.5) |
123 (14.9) |
|
|
Los Angeles |
334 (13.3) |
243 (14.4) |
91 (11.0) |
|
|
New York |
399 (15.9) |
310 (18.3) |
89 (10.8) |
|
|
Baltimore |
282 (11.2) |
204 (12.1) |
78 (9.5) |
|
|
Washington DC |
408 (16.2) |
272(16.1) |
136 (16.5) |
|
|
Pittsburgh |
96 (3.8) |
77 (4.6) |
19 (2.3) |
|
|
Calendar year |
|
|
|
0.089 |
|
2002-2006 |
781 (31.5) |
519 (31.1) |
262 (32.2) |
|
|
2007-2011 |
976 (39.3) |
680 (40.8) |
296 (36.4) |
|
|
2012-2017 |
724 (29.2) |
469 (28.1) |
255 (31.4) |
|
|
Average follow-up years, mean (SD) |
7.0 (3.7) |
7.1 (3.6) |
7.0 (3.9) |
0.478 |
|
Died during study |
904 (35.9) |
569 (33.7) |
335 (40.6) |
0.001 |
Reviewer 3 Report
Bahji A et. al investigates the association between cannabis use and HIV-1 RNA (viral load) among people with HIV (PWH). The finding suggested that among people living with HIV who are currently engaged in care and receiving ART, cannabis use is associated with decreased adherence in unadjusted analyses, but does not have direct impact on the viral control.
The main plus point of this study is high sample size.
However, few minor modification is required.
I would request to either plot one graphical representation or make a separate table of table no 1 subsection (HIV related factor vs cannabis use categories).
Check for few typo errors.
Author Response
Comment #1. I would request to either plot one graphical representation or make a separate table of the table with no one subsection (HIV-related factors vs. cannabis use categories).
Response #1. Thank you for this suggestion – we agree that figures are helpful! Per this feedback, we have created forest plots to summarize the unadjusted and adjusted models (Figure 2, corresponding to Table 2; and Supplementary Figure 1, corresponding to Supplementary Table 2).
Comment #2. Check for a few typo errors.
Response #2. We have edited the manuscript extensively for typos and grammatical errors.
