We thank Dr. Agrifoglio et al. for their comments, which come to complement and reinforce our conclusions.1 There is currently no discussion that the decision for admission to the ICU of patients infected with HIV, as in other patients with other chronic diseases, should be individualized and that a large percentage of these patients may benefit from intensive treatment, including all necessary support measures.

Although, apparently, the reasons for admission in our series seem different, both coincide in pointing out the importance of receiving or not receiving HAART at the time of admission, in addition to being aware of the infection or not, not to predict the prognosis, but as an aid to establish the diagnosis of the pathology that causes ICU admission. Agrifoglio et al. report that 70% of patients who were aware of their infection were admitted due to a non-AIDS-related pathology and we observed that sepsis caused 63% of ICU admissions in patients who were not receiving HAART.2

As we say, HIV infection has become a chronic disease, but that does not mean that its study within critical pathology has lost interest, as evidenced by two recent reviews published in high impact journals.3,4 Both summarize the differential diagnosis, based on the immunological status, of the main reasons for admission to the ICU of these patients: acute respiratory failure and altered consciousness, which can be challenging, especially in the most immunosuppressed patients. Another unresolved challenge they address is identifying the best time to start antiretroviral treatment, especially in patients who do not receive it and who are admitted to the ICU as a consequence of an opportunistic infection. In the reviews mentioned above, the authors give their expert recommendation and suggest that this is a possible area of research for the future.

Like Agrifoglio, the two reviews mention the importance of improving the knowledge of the outcome of these patients, they are more focused on the impact of ICU admission on the long-term outcome and Agrifoglio et al. is more focused on the selection of which patients will benefit from more aggressive measures. Even when we analyze the cohort studies published since 2005 (as Azoulay et al. did), we find that ICU mortality varies from 14 to 66%, a margin too wide to make decisions and inform the patient or family of the chances of survival. As Agrifoglio et al., who mention the combination of the stage of the HIV infection, the reason for admission and the accompanying medical factors to better predict the outcome, we believe that the combination of demographic variables, comorbidities, nutritional and immune status, reason for admission, and need for organic support in the first 72h of admission could be useful in developing a prognostic score that predicts hospital mortality.5