Article Text


Determinants of hospital mortality of HIV infected patients from north India
  1. S K Sharma,
  2. T Kadhiravan,
  3. A Banga,
  4. I Bhatia,
  5. T Goyal,
  6. P K Saha
  1. Department of Medicine, All India Institute of Medical Sciences, New Delhi, India
  1. Correspondence to:
 Dr S K Sharma
 Department of Medicine, Division of Pulmonary and Critical Care Medicine, All India Institute of Medical Sciences, New Delhi-110029, India;

Statistics from

A majority of the HIV infected population lives in developing nations. Most patients require hospitalisation for management of opportunistic infections (OIs) sometime during the course of their illness. Locally endemic infections and underlying malnutrition tend to influence the manifestations and course of the disease.1 However, there is paucity of data on pattern of disease and determinants of immediate outcome of such patients from Indian subcontinent.2

We report the determinants of hospital mortality in a cohort of 135 consecutive cases of HIV/AIDS, aged 13 years and above, admitted to the All India Institute of Medical Sciences (AIIMS), New Delhi, during the period of January 2000 through July 2003. These patients had been hospitalised for suspected OIs, and all patients underwent examination for diagnosis with subsequent management as per standard guidelines. For patients with Pneumocystis jiroveci pneumonia (PCP) whenever hypoxaemia was severe (PaO2<70 mm Hg; n = 5), corticosteroids were given in addition to oral co-trimoxazole. None of these patients received assisted ventilation. Secondary prophylaxis for the OIs was initiated as recommended.3

Mean age of the patients was 34 (SD 10) years and 23 patients (17%) were women. CD4+ cell counts were done in 109 patients. Most of these patients (82.6%) had CD4+ counts less than 200 cells ×106/l. Fifty patients (46%) had CD4+ counts less than 50 cells ×106/l. The mean number of OI was 1.4 per patient. The commonest OI was tuberculosis (TB) (71.1%), followed by oral candidiasis (39.3%). Other OIs (full data presented elsewhere) included PCP (n = 10), cryptococcal meningitis (n = 8), cerebral toxoplasmosis (n = 5), cytomegalovirus retinitis (n = 3), visceral leishmaniasis (n = 2), and progressive multifocal leucoencephalopathy (n = 1).

Twenty one patients (15.6%) died in hospital, most of them as a result of TB (n = 16; 76.2%) and PCP (n = 4; 19%). Factors associated with hospital mortality, on bivariate analysis, are shown in table 1. After adjusting for other factors (by multivariate logistic regression analysis), PCP was the only independent determinant and was associated with a more than fourfold increased risk of hospital mortality (adjusted odds ratio (95% CI): 4.7 (1.1 to 20.9); p = 0.041).

Table 1

 Predictors of in-hospital mortality in 135 HIV infected patients

Overall hospital mortality of 15.6% in this cohort is considerable and reflects the advanced nature of the disease at presentation. As our institute is a tertiary care facility and a national referral centre, this is expected. None the less, it may be possible that some OIs remained undiagnosed and indirectly influenced the outcome. This does occur as was shown in a necropsy study where it was found that a large number of potentially fatal OIs were not diagnosed antemortem.4

Unexpectedly, CD4+ counts had no independent effect on mortality. A similar observation has been reported in some previous studies.5,6 It appears that the virulence of the pathogen causing the OI, rather than the stage of the underlying disease, tends to influence the short term outcome. This finding has important therapeutic implications, especially because almost all these patients die of an OI.

It is suggested that any HIV infected patients with an OI, irrespective of the stage of the disease, should be managed with an aggressive approach. Once they recover from the OI, they can be offered antiretroviral therapy, which, over the years, has become extremely potent and effective. Such an approach is likely to improve the long term outcome of these patients.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.