Skip to main content

Four-year CD4 outcome estimates in advanced HIV: sensitivity to mortality handling, endpoint definitions and observation rules in records from Hengyang and three other Chinese centres

In brief

Accounting for deaths puts four-year CD4 recovery at 64%, not 79%

In adults with advanced HIV, the estimated chance of a documented CD4 count reaching 200 before death by four years was 64% when deaths were treated as competing events, versus 79% with a method that ignored them. Estimates varied from 44% to 74% across sensitivity analyses; the survivor-selected records cannot show outcomes for a true inception cohort.

Journal
International journal of infectious diseases : IJID : official publication of the International Society for Infectious Diseases (Q1)
Published
18 November 2026
Study design
Prospective / inception cohort
Evidence level
Level 2, Moderate (CEBM 2b)
Authors
Qianning Liu, Song Ou, Jing Ji, Chongxing Cao, Qingsong Shan, Xia Yin, et al.
PMID
43051518
DOI
10.1016/j.ijid.2026.109264

Why clinicians should know about it

  • Picked for Epidemiology (top studies of the week, 22 November 2026): Directly analyzes selection bias, mortality handling, and endpoint definitions

Abstract

Objectives To examine how mortality, observation and endpoint definition influence four-year CD4 outcome interpretation in routine advanced-HIV care. Methods We combined a Hengyang roster of patients alive and on ART at export with the centre's death records for adults with baseline CD4 <200 cells/µL, and described death timing in three further centres. We estimated first documented CD4 ≥200 before death by day 1,461 (Aalen-Johansen) and survival with Year-4-window CD4 ≥200, with sensitivity analyses. Results Providers reported that the roster was selected on survival, retention, suppression and regular CD4 testing; the death source held all known deaths. After the provider-recommended exclusion of roster patients treated under four years, 1,043/2,159 established composite outcomes among 2,750 adults were successful (48.3%); 591 were unresolved. Across four centres, 799/1,183 (67.5%) identified adult deaths occurred before four years. Among measured roster survivors, 902/1,243 (72.6%) had Year-4 CD4 ≥200. Using all dated measurements, recovery-before-death incidence was 64.3% (95% CI 62.4-66.2) versus naive 1-Kaplan-Meier 78.6%; sensitivity estimates ranged from 44.2% to 73.5%. Conclusions The measured-roster proportion summarises retained survivors. Mortality, observation schedules and endpoint definitions materially change CD4 summaries, and adding deaths to a survivor-selected roster does not reconstruct an inception cohort.

Abstract as published, via PubMed.

View on PubMedFull text at the publisherOpen in the app

For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.