Skip to main content
. 2020 Aug 29;73(7):e2005–e2015. doi: 10.1093/cid/ciaa1198

Table 4.

Associations with Coronavirus Disease 2019 Death from Cox Proportional Hazards Models

All public-sector SARS-CoV-2 cases diagnosed before 1 June 2020,a n = 15 203 Hospitalized public-sector SARS-CoV-2 cases, n = 2978
Adjusted HR 95% CI P value Adjusted HR 95% CI P value
Sex
 Female Ref Ref
 Male 1.45 1.22–1.74 <.001 1.29 1.09–1.53 .003
Age
 20–39 years Ref Ref
 40–49 years 3.19 2.06–4.93 <.001 1.83 1.23–2.72 .003
 50–59 years 10.84 7.34–16.01 <.001 3.81 2.68–5.42 <.001
 60–69 years 24.87 16.67–37.11 <.001 6.11 4.27–8.75 <.001
 ≥70 years 38.32 25.47–57.64 <.001 7.53 5.23–10.84 <.001
Diabetes
 None Ref Ref
 Diabetes HbA1c <7% 2.21 1.57–3.12 <.001 1.44 1.06–1.96 .020
 Diabetes HbA1c 7–8.9% 3.41 2.59–4.51 <.001 1.81 1.39–2.35 <.001
 Diabetes HbA1c ≥9% 3.62 2.85–4.59 <.001 1.60 1.27–2.0 <.001
 Diabetes, no HbA1c measurement 2.02 1.47–2.76 <.001 1.13 .83–1.55 <.001
Other noncommunicable diseases
 Hypertension 1.02 .84–1.24 .843 1.05 .88–1.27 .574
 Chronic kidney disease 1.92 1.51–2.45 <.001 1.51 1.20–1.89 <.001
 Chronic pulmonary disease/asthma .92 .72–1.18 .512 .68 .53–.86 .002
Tuberculosis
 Never tuberculosis Ref Ref
 Previous tuberculosis 1.55 1.19–2.02 .001 1.40 1.08–1.82 .011
 Current tuberculosis 1.62 1.04–2.51 .031 1.09 .72–1.65 .683
HIV
 Negative Ref Ref
 Positive 1.70 1.32–2.18 <.001 1.45 1.14–1.84 .002
  VL <1000 copies/ml (last 15 mo) and ART script (last 6 mo)b 1.60 1.19–2.17 .002 1.57 1.18–2.07 .002
VL <1000 copies/ml (2yr to 15 mo prior), or ART script (last 6 mo) and VL <1000 copies/ml >2yr prior 1.56 .80–3.07 .193 1.33 .72–2.46 .357
  VL ≥1000 copies/ml (last 15 mo) or CD4 count <200 cells/µl (last 18 mo) 3.39 1.14–3.62 <.001 1.60 .79–3.25 .190
  No VL (last 15 mo); CD4 count ≥200 cells/µl or unknown (last 18 mo) 1.73 1.10–2.71 .017 1.17 .73–1.87 .506
ART in PLWH with script issued in last 12 monthsc
 Abacavir or zidovudine Ref Ref
 Tenofovir disoproxil fumarate .41 .21–.78 .007 .57 .31–1.04 .067
 Efavirenz Ref Ref
 Lopinavir .91 .37–2.25 .846 .68 .29–1.63 .392
 Atazanavir .38 .05–2.92 .352 1.09 .25–4.82 .911
 Dolutegravir .57 .16–2.01 .380 .62 .17–2.22 .461
ART duration
 <1 year Ref Ref
 1–2 years .78 .21–2.94 .719 1.28 .37–4.42 .701
 ≥2 years .54 .19–1.48 .230 .55 .21–1.42 .213
CD4 count during COVID-19d
 >350 cells/µl 1.24 .95–1.63 .112
 200–349 cells/µl 1.65 .94–2.88 .080
 <200 cells/µl 2.36 1.47–3.78 <.001

Data are multivariate HRs and 95% CIs among (1) all adult COVID-19 cases diagnosed before 1 June 2020 (n = 15 203) and (2) all hospitalized adult COVID-19 cases (n = 2978).

Abbreviations: ART, antiretroviral therapy; CI, confidence interval; COVID-19, coronavirus disease 2019; HbA1c, glycosylated haemoglobin; HIV, human immunodeficiency virus; HR, hazard ratio; PLWH, people living with HIV; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; VL, viral load.

aAnalysis limited to cases diagnosed before 1 June 2020, when testing criteria changed, with public-sector tests being limited to patients >55 year of age or with comorbidities.

bReference category is HIV negative; adjusted for all other variables listed in this table in a model that included the listed categories of HIV VL, CD4 count, and ART instead of the binary variable of testing HIV positive vs negative; the effect of the other variables on mortality was similar to those presented here.

cRestricted to patients with documented antiretrovirals dispensed in the last 12 months, adjusted for all other variables listed in this table in a model that included the relevant antiretrovirals and ART duration; the effect of the other variables on mortality was similar to those presented here.

dReference category is HIV negative, restricted to patients living without HIV and 199 of 601 PLWH with a CD4 count measurement at the time of COVID-19 diagnosis or admission and adjusted for all other variables listed in this table in a model that included the listed categories of CD4 count instead of the binary variable of testing HIV positive vs negative; the effect of the other variables on mortality was similar to those presented here.