Background: Renal, cardiovascular, diabetic, and hepatic condition-specific hospitalization rates of HIV-infected women have not been examined in sufficient detail to evaluate trends over time.

Methods: The study included 885 HIV-infected women (median age 35, 75% had enrollment income ≤ $12,000 yr, 61% African American, 17% Hispanic, from the multicenter HIV Epidemiology Research Study [HERS]), who were followed for 4.9 years (mean) at 15 semi-annual research visits. Hospitalization diagnoses for AIDS-defining conditions and for renal, cardiovascular, diabetic, and hepatic conditions were abstracted from inpatient medical records. We calculated CD4⁺ cell count-adjusted rate ratios (RR) and p-values for year of hospitalization using Poisson regression with repeated measures and GEE estimation.

Results:

Poisson regression-adjusted# condition-specific hospitalization rate ratios vs. 1994
		1995		1996		1997		1998		1999
	n*	RR	p	RR	p	RR	p	RR	p	RR	p
Renal conditions	182	1.3	0.62	1.3	0.59	1.2	0.72	2.0	0.23	1.6	0.46
Cardiovascular conditions	360	1.3	0.28	1.7	0.06	1.8	0.02	2.1	<.01	2.0	0.02
Diabetes	108	1.8	0.38	1.4	0.42	0.9	0.81	2.7	0.06	1.8	0.32
Hepatic conditions	186	4.0	0.01	4.1	.008	3.9	0.02	8.0	<.01	10.2	<.01
AIDS-Defining	349	1.5	0.17	2.8	<.01	2.7	<.01	2.2	<.01	1.4	0.35

  * n of hospitalizations ^# adjusted for CD4+ cell count at each study visit

We found 192 women with pre-existing renal abnormalities (serum creatinine ³ 1.4 or urinary protein 2+ or greater) had increased rates of renal condition hospitalizations in 1998–1999 vs 1994–1997 (RR = 2.2, p = 0.02). But 693 women without such abnormalities did not have increased rates (RR = 0.9, 1998–1999 vs 1994–1997, p = 0.81).

Conclusions: In this cohort of mostly non-white, socio-economically disadvantaged women from four inner-city locations in the U.S., AIDS hospitalization rates peaked in 1996–1997. However, rates increased thereafter for cardiovascular conditions and especially so for hepatic conditions. Rates of renal condition-specific hospitalizations were flat overall but increased over time in the subgroup of women with pre-existing high creatininemia or proteinuria.