Table 1.
Cohort characteristics by ever exposure to efavirenz
| Variable | Ever-EFV * (n = 305) |
Only-NVP (n = 389) |
P-value |
|---|---|---|---|
| Demographic Characteristics | |||
| Age, median (IQR) | 32 (27, 40) | 34 (28, 39) | 0.165 |
| Female, n (%) | 201 (65.9%) | 283 (72.8%) | 0.051 |
| Baseline pregnant | 48 (25.3%) | 41 (14.9%) | 0.005 |
| Married, n (%) | 169 (55.4%) | 179 (46.0%) | 0.014 |
| Educational attainment, n (%) | 0.051 | ||
| None | 54 (17.7%) | 53 (13.6%) | |
| Some primary | 95 (31.1%) | 156 (40.1%) | |
| Completed primary | 64 (21.0%) | 85 (21.9%) | |
| Some secondary and higher | 92 (30.2%) | 95 (24.4%) | |
| Asset index†, n (%) | 0.014 | ||
| 1st quintile (most poor) | 46 (15.1%) | 90 (23.1%) | |
| 2nd quintile | 57 (18.7%) | 83 (21.3%) | |
| 3rd quintile | 60 (19.7%) | 77 (19.8%) | |
| 4th quintile | 69 (22.6%) | 77 (19.8%) | |
| 5th quintile (least poor) | 73 (23.9%) | 62 (15.9%) | |
| Year of enrollment, n (%) | < 0.001 | ||
| 2005–07 | 68 (22.3%) | 251 (64.5%) | |
| 2008–10 | 32 (10.5%) | 103 (26.5%) | |
| 2011–13 | 205 (67.2%) | 35 (9.0%) | |
| Clinical Characteristics | |||
| Efavirenz use at enrollment, n (%) | 253 (83.0%) | 9 (2.3%) | < 0.001 |
| NRTI backbone at enrollment, n (%) | < 0.001 | ||
| 3TC + TDF | 172 | 11 | |
| 3TC + AZT | 117 | 285 | |
| FTC + TDF | 1 | 1 | |
| 3TC + D4T | 15 | 92 | |
| Probable depression at enrollment‡, n (%) | 99 (32.5%) | 129 (33.2%) | 0.84 |
| Suicidal ideation at enrollment, n (%) | 21 (6.9%) | 26 (6.7%) | 0.92 |
| Enrollment CD4+ count (cells/μL), median (IQR) | 231 (126, 348) | 151 (88, 219) | < 0.001 |
| Enrollment viral load (log10 copies/mL), median (IQR) | 0.83 (0.67, 1.00) | 0.86 (0.67, 1.00) | 0.75 |
| Tuberculosis co-infection | 26 (8.5%) | 8 (2.1%) | < 0.001 |
| Physical Health Summary score§, median (IQR) | 57.3 (48.6, 60.8) | 54.7 (46.2, 58.9) | 0.001 |
| Heavy drinking∥, n (%) | 49 (16.7%) | 74 (19.9%) | 0.28 |
| Follow-up Characteristics | |||
| Duration of follow-up (weeks), median (IQR) | 96 (95, 97) | 96 (95, 97) | 0.163 |
| Number of follow-up visits, median (IQR) | 6 (6, 6) | 7 (6, 8) | < 0.001 |
| Number of follow-up visits without missing covariate data, median (IQR) | 6 (4, 6) | 7 (5, 7) | < 0.001 |
| Proportion of follow-up visits with viral suppression¶, median (IQR) | 0.83 (0.67, 1.00) | 0.86 (0.67, 1.00) | 0.75 |
| Died, n (%) | 6 (2.0%) | 11 (2.8%) | 0.47 |
| Lost to follow-up**, n (%) | 13 (4.3%) | 40 (10.3%) | < 0.001 |
| Total number of missing visits, n (%) | 162 (10.6%) | 423 (18.1%) | < 0.001 |
| Number of missing visits per participant††, n (%) | 0 (0, 1) | 1 (0, 2) | < 0.001 |
| Number of visits censored for missing covariate data per participant, n (%) | 0 (0, 1) | 0 (0, 0) | <0.001 |
EFV = efavirenz; NVP = nevirapine; IQR = interquartile range; NRTI = nucleoside reverse transcriptase inhibitor; 3TC = lamivudine; TDF = tenofovir disoproxil; AZT = zidovudine; FTC = emtricitabine; D4T = stavudine
The ever-efavirenz group included participants who met criteria for efavirenz exposure for at least one follow-up visit in study period; the only-nevirapine group included participants who did not meet criteria for efavirenz exposure in any follow-up visits in study period
Asset index was calculated based on the methods described by Filmer and Pritchett (21).
Probable depression was defined as a score of >1.75 on the Hopkins Symptom Checklist-15 modified for the local Ugandan context (14–16,19,20).
The Physical Health Summary score was estimated using the Medical Outcomes Survey-HIV score (22).
Heavy drinking was based on the 3-item heavy consumption subset of the Alcohol Use Disorders Identification Test - Consumption screen (23).
Viral load suppression was defined by an undetectable viral load at the nearest visit after 2 months of antiretroviral therapy. The limit of detection ranged from 400 to 20 copies/mL as the study progressed, depending on the assay available at the time of the study visit.
Lost to follow-up was defined as not having a study visit 550 days after antiretroviral therapy initiation or having withdrawn from the study before 2 years.
The number of missed visits before the last recorded study visit. 1 missed visit was added if the last study visit was >3.5 months before the two-year time endpoint of the analysis.