Bacteria can infect the lungs in people with HIV, especially after the CD4 count falls below 200 cells/mm³. Smokers are more vulnerable to bacterial lung infections. These can cause either a mild cough or a severe pneumonia. This pneumonia can be extremely difficult to distinguish from PCP. A bronchoscopy may be necessary to diagnose bacterial pneumonia.

People taking T-20 (enfuvirtide, Fuzeon) seem to be at slightly increased risk of bacterial pneumonia. Studies suggest that if 100 people take T-20 for a year, four or five will develop bacterial pneumonia. Experts believe this increased risk of bacterial pneumonia is occurring because T-20 is injected. See T-20 - overview in Drugs used by people with HIV: Entry and fusion inhibitors for details.

As with other opportunistic infections, the incidence of bacterial pneumonia has fallen since combination antiretroviral therapy was introduced, such that patients treated with protease inhibitors have similar rates of bacterial pneumonia to HIV-negative patients (Le Moing 2006).

When bacterial pneumonia does occur now, it is associated with a relatively poor prognosis. A study of bacterial pneumonia in Barcelona, Spain, found that the yearly rate of bacterial pneumonia between 1997 and 2002 was eight per 1000 HIV patients^[1]. In the era of combination therapy, people who develop bacterial pneumonia tend to have high viral load, or advanced HIV disease and other conditions such as cirrhosis^[2]. Prognosis was poorer than before HAART.

Between 1995 and 1998, 9% of HIV admissions to one clinic were due to bacterial lung infections and this group often had more severe illness and longer hospital stays than individuals with other HIV-related illnesses^[3]. Injecting drug users tend to be at increased risk of bacterial pneumonia^[4], ^[5]. Such infection is more common in smokers than non-smokers. Other risk factors for bacterial pneumonia include breathing chemical irritants and hospitalisation for pneumonia^[6].

However, more recent findings have suggested that being HIV-positive does not lead to a worse outcome following hospitalisation for bacterial pneumonia^[7]. The length of hospitalisation and mortality rates were similar in 58 HIV-positive and 178 HIV-negative patients, although the HIV-positive patients' CD4 cell counts were mostly below 200 cells/mm³.

Streptococcus pneumoniae and Haemophilus influenzae are common bacterial causes of lung infection (pneumococcal pneumonia), as well as other bacteria in the streptococci family. In a Spanish study, a type of pneumococcal pneumonia was the first HIV-related illness in 48% of people with HIV^[8]. However, recent studies have shown that despite the fall in incidence since the introduction of HAART, the rate of pneumococcal disease remains elevated in HIV-positive patients, and at equal levels in patients with CD4 cell counts above and below 200 cells/mm³, as well as in patients with high CD4 cell counts, particularly smokers, injecting drug users, older patients and those with high viral loads due to poor adherence^[9], ^[10].

Even in the HAART era, patients with AIDS are 35 times more likely to become ill with invasive pneumonia than HIV-negative people. Invasive pneumonia occurs when the bacteria invade normally sterile sites, like the blood or cerebrospinal fluid^[11]. 

The bacterium Pseudomonas aeruginosa has also been reported as an increasingly common cause of cases of bacterial pneumonia in people with HIV^[12].

Pneumococcal infections usually respond to antibiotics such as penicillin, ciprofloxacin (Ciproxin), amoxicillin (Amoxin / Amix / Amoxident / Amoram / Galenamox / Rimoxallin) or antibiotics belonging to the macrolide or second-generation cephalosporin families. If the condition does not respond, this may indicate the presence of a concurrent infection such as PCP. Co-trimoxazole, which is effective against PCP as well as pneumococcal pneumonia, may be recommended if the symptoms are particularly severe.

Bacterial pneumonia caused by rarer organisms, such as Rhodococcus equi, Nocardia or Bordetella may require intensive, prolonged therapy with antibiotics.

A significant minority of people experience recurrent bacterial pneumonia. If tests show levels of immunoglobulin to be low, intravenous immunoglobulin may be given, although this has only be proven effective among children with CD4 counts greater than 200 cells/mm³.

Co-trimoxazole effectively prevents many bacterial infections, and prophylaxis in children with HIV in Zambia has been shown to reduce the rate of death in a major two year study, chiefly by reducing the incidence of bacterial pneumonia^[13].

Doctors may recommend an immunisation against S. pneumoniae and H. influenzae, as well as the standard influenza immunisation. The pneumococcal vaccine was approved without efficacy trials, based on evidence that the vaccine stimulated immune response. However, a recent study has raised serious doubts about the vaccine, since it did not affect the number of cases of pneumococcal disease and was associated with an increased rate of overall cases of pneumonia when it was tested in Uganda^[14]. The researchers suggested that while the vaccine might be effective in adults in Europe or North America, it was probably ineffective among Ugandan adults because the strains of pneumococcus used to develop the vaccine were different from those which are found in Uganda.

Streptococcal pneumonia may also be prevented by co-trimoxazole prophylaxis^[15], providing that co-trimoxazole resistance is not widespread in the local population. If resistance is present, this means that the streptococcal bacteria circulating in the area will have some degree of resistance to co-trimoxazole, and the widespread use of co-trimoxazole prophylaxis may actually encourage the development of further resistance and the loss of protection from this useful antibiotic. See Co-trimoxazole in Drugs used by people with HIV: Antibiotic drugs for further details.