Feasibility of primary HIV care in a Thai tertiary care centre

Sir: Data exist to support the effectiveness of primary care for HIV type-1 (HIV-1)-infected patients via comprehensive HIV treatment and prevention in developed countries. ^1,2 However, adoption or modification of these strategies in resource-limited settings (RLSs) have not been well studied. Variations in demographic characteristics and incidence of opportunistic infections (OIs) by geographic region also have implications for modified HIV tiered-screening strategies in RLSs. ^3,4

We conducted a study of HIV primary care (HPC) among new antiretroviral therapy (ART)-naïve adults (>15-year-old) attending the clinic at a Thai tertiary care centre. Eligibility required HPC patients to be enrolled in 2004 and prospectively followed for three years. At the initial visit, all patients received baseline screening for CD4 count, hepatitis A (anti-HAV), B (HBsAg, anti-HBc IgG, anti-HBsAb), C (anti-HCV), tuberculosis (TB) via a two-step tuberculin skin test (TST) and a chest radiography (CXR), Venereal Disease Research Laboratory (VDRL) test with subsequent confirmatory treponemal test (if needed) and cervical Papanicolaou (PAP) smears. Hepatitis A and B immunizations were given if indicated. Patients with reactive TST and negative CXR were treated for latent TB, while those with abnormal CXR were investigated for active pulmonary TB and treated accordingly. ⁵ At the second visit, tiered-screening occurred by CD4 count: (1) patients with CD4 count 100–199 cells/µL (category [CAT]I); (2) patients with CD4 count 50–99 cells/µL (CATII); and (3) patients with CD4 count <50 cells/µL (CATIII). CATII and III patients underwent toxoplasma serology and serum cryptococcal antigen (SCRAG) screening prior to receipt of anticryptococcal chemoprophylaxis. ⁶ CATIII patients underwent examination for cytomegalovirus (CMV) retinitis. CATI, II and III patients received ART, pneumocystis (PCP) and anticryptococcal chemoprophylaxis. ⁶ Follow-up scheduling for clinical and laboratory assessment were according to our clinic protocol. ⁵ The primary outcome was the incidence of TB, PCP, cryptococcosis, CMV retinitis and toxoplasmosis in HPC patients compared with Thai national patients (NPs) from the Ministry of Public Health HIV database during the same follow-up period. ⁷ The NPs had similar demographics and HIV risk factors as HPC patients but did not receive any tiered-screening.

There were 257 HPC patients; mean age 36 years (range, 15–61 years), 155 (60%) men, median CD4 count 147 cells/µL (range, 0–957 cells/µL), sexual contact is the major HIV risk (97%). There were 312,388 NPs; mean age 34 years (range 15–70 years), 218,029 (70%) men, median CD4 count was 41 cells/µL (range, 13–113 cells/µL), sexual contact (84%) is the major HIV risk. ⁷ At the initial visit, 46 (18%) HPC patients were CATI, 56 (22%) were CATII, 88 (34%) were CATIII, 132 (51%) had positive anti-HAV, 29 (11%) had chronic active hepatitis B, 12 (5%) had positive anti-HCV, two (1%) had serological evidence of syphilis, 38 (15%) with reactive TST and 41 (16%) with abnormal CXR. Among these 41 patients, all had confirmed active pulmonary TB. Notably, 14 (34%) had non-reactive TST. Six of 102 women (6%) had abnormal PAP smears and underwent colposcopy. Three (3%) had cervical cancer. Among 144 patients in CATII and III, nine (6%) had positive SCRAG, nine (6%) had positive toxoplasma IgG; all had no evidence of infections. Six (3%) of 88 CATIII patients had CMV retinitis. During the follow-up period, one patient each from CAT II developed cryptococcal meningitis and cerebral toxoplasmosis due to non-compliance to chemoprophylaxis. After adjustment for CAT level, the three-year incidences of TB, PCP, cryptococcosis, CMV retinitis and toxoplasmosis in HPC patients were significantly lower than those of NPs (Table 1). Numbers needed to treat were highest for TB, followed by PCP and cryptococcosis.

OPEN IN VIEWER

Table 1

Incidence of HIV-related infections in HIV primary care and Thai national patients during the three-year follow-up period 2004–2006

Infections	HIV primary care patient*†	Thai national patient*	P value‡§	NNT**
Tuberculosis	0	499	<0.001	20
Pneumocystis	0	313	<0.001	32
Cryptococcosis	39	144	<0.001	95
CMV retinitis	0	67	<0.001	149
Toxoplasmosis	39	48	<0.001	1111

*Rate/10,000 patients

^†The routine schedule was sequentially once every two weeks for four visits, once every four weeks for four visits, once every two months for three visits, and then every three months ⁵

^‡HIV primary care patient versus Thai national patient

^§Incidences were compared using the chi-square test. All P values were two-tailed; P < 0.05 was considered statistically significant

**Calculated inversion of differential rates of indicator infections between HIV primary care patient and Thai national HIV patient

CMV = cytomegalovirus; NNT = number needed to treat

Our findings have notable clinical implications. Firstly, in RLSs with a high prevalence of TB, two-step TST is not a sensitive screening test in HIV-infected patients. Notably, 14 active pulmonary TB cases with non-reactive TST were identified by routine CXR screening. This was consistent with findings from a previous study that a subset of TST non-reactive HIV-infected patients with risk factors for TB exposure, should proceed with screening CXR. ⁸ Secondly, we identified a significant number of cryptococcal antigenaemia without meningitis. Our screening was shown to be highly effective for identifying patients at risk of cryptococcal meningitis that required additional work-up prior to receipt of anti-cryptococcal chemoprophylaxis. ^6,9,10 Lastly, HPC is feasible and effective in prevention of OIs in this RLSs. Additional studies are needed to assess cost-effectiveness of HPC and appropriate screening tests for HIV-infected persons in RLSs.