Pneumocystis carinii pneumonia (PCP) was one of the first identified manifestations of AIDS. Although improved strategies for prevention and treatment of PCP have greatly reduced the number of cases and deaths, PCP and other pneumonias remain a major cause of illness and death among people with AIDS, and account for roughly one-third of AIDS deaths. Because bacterial pneumonias and other respiratory problems can sometimes produce similar symptoms, diagnosis of PCP can be tricky. Today, use of prophylactic treatment to prevent PCP in persons with fewer than 200 CD4 cells/mm3 is universally regarded as one of the most effective medical interventions available.
History and Trends
Pneumonia -- infection and inflammation of the lungs in which parts of the lungs may fill with fluid -- can be caused by a wide variety of organisms, including bacteria, viruses and fungi. Whatever the cause, pneumonia can be debilitating and often deadly.
Until the 1980s pneumonia caused by Pneumocystis carinii was extremely rare: only about 50 cases were reported in 1970. Although the organism--long regarded as a protozoan, but now understood to be more like a fungus--is quite common, it is harmless in individuals with healthy immune systems. Before AIDS, the few instances of PCP disease occurred in people whose immune systems had been seriously weakened. Cases were reported in malnourished World War II orphans and, more recently, in organ transplant recipients receiving immunosuppressive anti-rejection therapy and in people with cancer undergoing chemotherapy. When PCP began appearing in 1980 and 1981 in young gay men with none of the usual risk factors, alarms went off at the Centers for Disease Control (CDC).
In the early 1980s, doctors had little experience treating PCP; the death toll was astonishing, and an average patient had only a 50% chance of surviving an episode of the disease. Among those whose illness was severe enough to require a breathing tube and ventilatory support, the incidence of death reached 80%. For many years, PCP remained the number one killer of people with AIDS. Even those who recovered from a first episode of PCP generally saw their health continue to decline. In over 60% of cases, PCP recurred, often with increasing severity. Despite recent progress, PCP is still cause for concern. As San Francisco physician Lisa Capaldini, MD, put it recently, "If someone gets PCP out of the blue, they're likely to have a progressive decline in their health, because a rip-roaring pneumonia can really knock your socks off."
Fortunately, progress on all fronts against PCP has made the outlook more hopeful since the late 1980s. Improved diagnostic and treatment methods began to increase survival. Perhaps most importantly, increasing use of prophylactic medications to prevent PCP began to reduce the number of cases.
An analysis by CDC researchers of HIV-related causes of death listed on death certificates from 1987-1992 illustrates the point dramatically. While the rates of death involving most other opportunistic infections remained fairly constant, PCP-related mortality dropped from 32.5% of HIV-related deaths in 1987 to 13.8% by 1992. In that same period, deaths due to bacterial pneumonia increased from 1.2% to 2.1%. Unspecified pneumonias remained constant at about 18%. Some of these unspecified pneumonias no doubt represent undiagnosed cases of PCP, while others represent divergent causes. The researchers believe the decline in PCP canceled out increases in other types of pneumonia in this "unidentified" group.
A similar trend has been seen in another CDC study, the Adult Spectrum of Disease Project, which collected data from 100 inpatient and outpatient medical facilities in 10 U.S. cities from 1990-1995. Although only some of the data have been released, figures through 1993 show a gradual but steady decline in PCP incidence among gay and bisexual men. In contrast, PCP rates increased among injection drug users (IDU), although the increase did not quite achieve statistical significance. Fewer of the IDU were receiving PCP prophylaxis, and the researchers theorize that those who were may not have adhered to their prophylactic regimens as well as the gay and bisexual men. Data on other causes of pneumonia in this study have not yet been released.
The incidence of PCP in San Francisco has more or less conformed to the national trend. Annual PCP cases peaked at 1,095 in 1987, leveled off for a few years, and dropped to 613 by 1994. Data for 1995 are still incomplete. According to Kevin McKinney of the city's AIDS Office, reliable figures are not available for other types of pneumonia.
Still, for a city as renowned for state-of-the-art HIV care as San Francisco, PCP is far from vanquished. John Stansell, MD, Medical Director of the AIDS Program at San Francisco General Hospital, notes, "We still see a very large number of PCP cases here. Per month it's probably gone from a high of about 30 cases to about 20-25. It's true that there's been somewhat of a decline, but it's by far still the most frequent opportunistic infection that we encounter here at San Francisco General."
About half of those cases, Stansell continues, occur in people who were not receiving PCP prophylaxis. Most of the others occurred in persons who did not take their medication as scheduled or who were on aerosolized pentamidine or dapsone, which are considered weaker PCP preventatives than trimethoprim-sulfamethoxazole (TMP-SMX, brand names Bactrim and Septra). Stansell believes that true "breakthrough" infections while on TMP-SMX are "very rare" (see below for a detailed discussion of PCP prophylaxis).
Stansell is part of the team conducting the long-term Pulmonary Complications of HIV Infection study, which recruited patients at 6 sites around the country beginning in 1987 and followed them for 5 years. The study enrolled 1130 HIV positive participants and 167 HIV negative controls matched for age and gender. Stansell calls it "the first study that really aimed at enrolling a cohort that looked like the AIDS epidemic" in terms of demographics. Much of the data are still awaiting publication, but some appeared last September in the New England Journal of Medicine (333(13): 845-51. September 28, 1995).
"The incidence of bacterial pneumonia was about 10-fold higher among HIV-infected people than it was in uninfected individuals," Stansell says. Although the rates of bacterial pneumonia--caused by various organisms including Streptococcus pneumoniae and Haemophilus influenzae--increased as CD4 counts declined, HIV positive persons with normal CD4 counts still had significantly higher rates than the uninfected controls. The explanation, Stansell suspects, is that "there are probably qualitative differences in the ability of the body to rid itself of pathogens if you're HIV-infected. The immune cells, even if they're there, probably don't work as well."
Smoking was also associated with higher rates of bacterial pneumonia. Strikingly, PCP prophylaxis with TMP-SMX was associated with a two-thirds reduction in bacterial pneumonia, in addition to a decrease in PCP.
Overall, those who experienced any type of pneumonia had a mortality rate 4 times higher than those who did not. Equally distressing, Stansell notes, is that "the people who have bacterial pneumonia appear to have a greater risk of rapid progression in their HIV disease. Those data are just being ferreted out now, but it looks like it's a very bad prognostic indicator."
PCP prevention has been one of the areas of greatest progress in HIV care over the years. Some precautions can be taken to reduce the risk of bacterial pneumonias as well.
The first line of defense, and one not to be underestimated, is to see a physician who is knowledgeable about HIV disease. An assessment of HIV care provided by primary care doctors published last year in Archives of Internal Medicine found that large numbers of physicians did not prescribe or recommend the most basic elements of HIV care. For example, only half indicated that they would start appropriate PCP prophylaxis.
If there is anything that AIDS-experienced physicians agree upon, it is that PCP prophylaxis works. Standard practice is to recommend prophylaxis for any HIV-infected person with 200 or fewer CD4 cells/mm3 or whose CD4 percentage is less than 14% of total lymphocytes. PCP prophylaxis is also recommended for anyone, regardless of CD4 count, who is experiencing a sharp decline in CD4 counts or persistent, unexplained fevers or thrush (shown in some studies to be independent predictors of high risk for PCP). Anyone who has already had 1 or more bouts of PCP should also receive prophylaxis. Study after study has demonstrated that effective prophylaxis sharply reduces the incidence of PCP in such individuals.
The CDC has developed specific PCP prophylaxis guidelines for infants and children. The agency recommends prophylaxis for all infants born to HIV-infected mothers starting at age 4-6 weeks, to be discontinued if HIV antibody testing after 4 months of age indicates that the child is HIV-uninfected. For HIV-infected children aged 1-5 years, prophylaxis is recommended for those with fewer than 500 CD4 cells/mm3 or a CD4 percentage less than 15%. TMP-SMX, an oral antibiotic, is the preferred drug. (For more on pediatric HIV infection, see HIV/AIDS in Children, this issue.)
TMP-SMX is the prophylactic drug of choice for adults as well. Other regimens, notably dapsone, another oral drug, and aerosolized pentamidine (AP), inhaled in mist form through a device called a nebulizer, are also used extensively. Besides being extremely effective at preventing PCP, TMP-SMX has 2 other advantages: it is inexpensive, and it also helps to prevent other diseases such as toxoplasmosis (a brain infection) and bacterial infections including pneumonia. However, TMP-SMX can produce a number of toxic side effects, including rash, fever, nausea, vomiting and neutropenia (loss of white blood cells). Life-threatening reactions, including severe anaphylactic reactions or Stevens-Johnson syndrome (a condition involving severe blistering and sloughing of skin and mucous membranes) can occur, but are rare. Aerosolized pentamidine does not prevent extrapulmonary PCP.
The scales were definitively tipped towards TMP-SMX by a study known as ACTG 021, in which 1 double-strength tablet of TMP-SMX daily was compared to 300 mg of AP given once per month to HIV positive persons who had already had PCP at least once. PCP recurred in the AP group more than twice as frequently as in the TMP-SMX group. Based on this study, the U.S. Public Health Service recommended this dose of TMP-SMX for PCP prophylaxis.
A later study, ACTG 081, published in the March 16, 1995 New England Journal of Medicine, seemed to muddy the picture a bit. This trial compared a higher dose of TMP-SMX (1 double-strength tablet twice daily) to dapsone (50 mg twice daily) to AP (300 mg every 4 weeks) in people with fewer that 200 CD4 cells/mm3 and no prior history of PCP. Those who had serious (grade 3) toxic effects had their doses reduced and/or were switched to a different therapy. Although TMP-SMX was more active against PCP than the other drugs, toxicity forced 79% of those receiving it to reduce the dose or switch to another drug.
The "intent-to-treat" analysis--comparing the groups based on the drug they were originally assigned, regardless of later changes--showed only statistically insignificant differences in the risk of developing PCP among the 3 groups. Over 36 months, the risk was 18% in the dapsone group, 17% in the TMP-SMX group and 21% in the AP group. For those with fewer than 100 CD4 cells/mm3, though, the difference was more pronounced: the risks of developing PCP were 22%, 19% and 33%, respectively.
However, an "as-treated" analysis--looking at the therapy the patient was actually receiving at the time of the PCP episode--tells a somewhat different story. Only 4 of 34 treatment failures in the TMP-SMX group occurred among people still taking the drug at the original dose, and none occurred in those whose dose was reduced to once daily. The failures overwhelmingly occurred after patients had switched drugs. By contrast, 21 of 33 dapsone treatment failures occurred while on that drug, and the failure rate increased 4-fold at the first dose reduction. Few AP patients switched drugs, and 37 of 38 failures occurred while on AP.
The researchers concluded that although TMP-SMX is clearly stronger against Pneumocystis (P.) carinii, use of this drug as the first prophylactic strategy did not produce significantly better results than starting with dapsone or AP. The reason, the researchers write, is that "because of continuing cross-overs and dose reductions as the trial progressed, more patients were receiving less adequate treatments even as their vulnerability to P. carinii increased" due to declining CD4 counts.
The impact of this study is somewhat uncertain. Clearly 2 double-strength TMP-SMX tablets per day showed no advantage over one. Just as clearly, 100 mg per day of dapsone per day performed better than 50 mg. Overall the researchers found "similar effectiveness," no matter which drug patients started with. An accompanying editorial went further, suggesting that AP, with its low toxicity, might be the drug of choice for those with CD4 counts over 100 cells/mm3, and TMP-SMX the preferred prophylaxis for those with fewer than 100 cells/mm3.
That suggestion goes against the generally accepted recommendation to use TMP-SMX as first-line prophylaxis across the board. There seems not to have been a general move in favor of AP since the study was published. A meta-analysis of 35 randomized PCP prophylaxis trials that appeared in the January 22, 1996 Archives of Internal Medicine sheds more light on the situation. The researchers pooled and analyzed the results from all 35 studies, which looked at a variety of drugs--including TMP-SMX, AP and dapsone--and dosing regimens. Some of the studies combined dapsone with pyrimethamine.
Such an analysis has limitations. In comparing results from different trials, it is impossible to completely adjust for all of the differences in trial methodology. Still, the researchers found clear patterns that allowed them to draw some fairly solid conclusions.
"Regardless of dose," they wrote, "TMP-SMX was almost universally effective for patients who tolerated it" [emphasis added]. Overall, the drug was 42% more effective than AP in preventing PCP. The difference between TMP-SMX and dapsone was not statistically significant, though the trend favored TMP-SMX. A dose of 100 mg of dapsone was more effective than 50 mg. Overall, AP had the fewest side effects while TMP-SMX had the most.
There were few treatment failures on TMP-SMX, most of which occurred after the drug had been discontinued. Most striking were the dose-response relationships uncovered. For example, increasing the daily dose of TMP-SMX seemed to hurt more than it helped; in fact, there was a trend toward more failures at the highest dose tried (2 double-strength tablets per day, the regimen used in ACTG 081). AP, meanwhile, did better when given twice monthly instead of the standard once monthly dose.
Thus, reduced dosing of TMP-SMX--such as 1 double-strength tablet 3 times per week--might minimize the need to discontinue therapy due to toxicity, with little loss in efficacy. The scientists caution that relatively little data exists to support using this regimen in people with lower CD4 counts or those who have already had PCP.
Another approach to dealing with the toxicity of TMP-SMX is known as desensitization. This involves starting with tiny, diluted doses of the drug and building up to the regular dose over a period of days, so that the body gradually learns to tolerate it. Many different desensitization protocols exist. One version offered by Project Inform takes 8 days; others take more or less time. Considerable success has been reported with desensitization, although it is generally considered too risky to attempt with persons who have had a severe anaphylactic reaction to sulfa drugs or who have had Stevens-Johnson syndrome.
What is less clear is which protocol is best. "There's no standard," observes San Francisco clinician and AIDS researcher Gifford Leoung, MD. "It's all arbitrary; there are no published data from well-organized trials. In fact," he continues, "there's a paper that came out last year that suggested you don't even need to desensitize people." Instead, you can "just give them a single dose--directly challenge them right up-front." Leoung is protocol chair of an American Foundation for AIDS Research-financed trial that is now comparing this direct-rechallenge approach with gradual desensitization.
In any case, Stansell declares, "The most important thing you can do for yourself is to take PCP prophylaxis." Whatever treatment you're on, it is critical to do it right. That means taking your pills as scheduled, and if side effects make that difficult, talking with your doctor about ways to manage the problem, rather than haphazardly skipping doses. If you are using AP, follow the instructions you are given about posture and breathing during the treatment. Slouching and breathing shallowly will prevent the medication from getting all the way into the lungs, reducing the benefit.
For protection against bacterial pneumonia, the CDC recommends vaccination with pneumococcal vaccine as soon as possible after testing HIV positive, a recommendation the agency calls "especially pertinent in light of the increasing incidence of invasive infections with drug-resistant strains of Streptococcus pneumoniae." Vaccination against Haemophilus influenzae "may be considered, but data are insufficient to recommend the use of this vaccine in HIV-infected adults at this time."
As Stansell points out, TMP-SMX seems to offer some protection against bacterial pneumonias as well as PCP. Although the CDC suggests this be considered when choosing PCP prophylaxis, it warns that "indiscriminate use of this drug 'when not indicated for PCP prophylaxis or other specific reasons' may promote the development of resistant organisms." In addition, the drug will not protect against drug-resistant strains.
A constellation of symptoms, including fever, cough, fatigue and shortness of breath, can signal the onset of pneumonia. Some or all of these symptoms may be present. Of course, many other conditions, such as bronchitis and the common cold, can produce some of the same effects. If symptoms continue beyond a few days or worsen, it is essential to seek medical evaluation. Caught early, most pneumonias can be successfully treated, often on an outpatient basis, but if allowed to worsen, they can be life-threatening.
It is important to be able to describe your symptoms in as much detail as possible, including the length of time they took to appear, because these details will help your physician to narrow down a diagnosis. Symptoms can be clues to the cause of the pneumonia. For example, bacterial pneumonias tend to develop rapidly, often within 24 hours or less, while other lung infections, including Aspergillus or Mycoplasma infections, tend to have a slower onset. PCP also tends to come on slowly. Leoung says symptoms may be barely perceptible at first but then worsen slowly over a period of weeks.
The cough can also be an important clue. A classic PCP cough is dry, with little or no sputum; any sputum that is produced is likely to be clear. Bronchitis, on the other hand, can often produce large amounts of sputum; and bacterial pneumonias produce greenish, rust-colored, bloody or yellow-tinged sputum.
Often a doctor will listen to the patient's chest with a stethoscope. Pneumonias often produce a characteristic breathing sound called "rales," while bronchitis usually does not; PCP may or may not produce recognizable chest sounds.
Chest X-rays are often used to make a diagnosis because different illnesses produce characteristically different patterns. However, a certain amount of natural variation occurs on X-rays, and PCP may not always be visible. Further testing is required to confirm a diagnosis made by X-ray.
Along with identifying the cause of the problem, Leoung notes, it is important to quickly make other key determinations. "How sick is this person? Do they need to be in the hospital? That's the real next question. If they're not getting enough oxygen they can't go home. Is this illness serious enough to require IV antibiotics, as opposed to oral antibiotics? If so, the person most likely needs to be hospitalized."
One important measure of the seriousness of any type of pneumonia is its effect on the amount of oxygen that gets into the blood. Two different tests are commonly used for making this determination. The simplest but crudest is known as pulse oximetry, which involves placing a device that resembles a clothes pin on a finger to measure the amount of oxygen under the skin.
Although oximetry is simple and fast, Stansell dismisses it as "very inexact." A far more precise method is to draw arterial blood and measure the gases it contains. Unfortunately, arterial blood gas measurement can be painful, because arteries tend to have more nerve endings than the veins used for most other blood tests.
Aside from these test results, the decision to hospitalize or not can involve what Leoung calls a somewhat subjective evaluation of "how much you think this person can tolerate in terms of getting worse. If there's a lot of leeway, you can go home because there's time to get you back in here. But if you're close to the edge, I don't want to send you home. Sometimes I rely on my gut feelings; the numbers may look great, but the person just does not look good." As with prophylaxis, these judgment calls tend to be most reliably made by an HIV-experienced physician.
If the symptoms, history and X-rays seem to point in one direction--PCP, for example--the physician will probably recommend beginning treatment immediately. Even so, standard practice in San Francisco is to make a confirmatory test for the responsible organism. Unusual symptomology can occur, and sometimes more than one condition may be present at once, with one set of symptoms partly masking another.
The least invasive way to confirm a diagnosis is through a procedure called sputum induction. The patient breathes a saline mist which provokes coughing, hopefully bringing up sputum which can be examined for P. carinii, bacteria, fungi or whatever organisms the symptoms and X-rays suggest. Patients are advised not to eat for 8 hours before the procedure, nor to use toothpaste or mouthwash, all of which can interfere with the results. Stansell calls sputum induction "a technically difficult procedure," which not everyone manages to do well. At San Francisco General Hospital, where the procedure is frequently used, over 70% of PCP diagnoses are made this way, but facilities that use the technique less are not usually this successful.
If sputum induction fails to enable a diagnosis, the next, more invasive step is bronchoscopy. In this procedure, a tube is inserted through the mouth or nose into the breathing tubes. The patient is generally sedated and the throat numbed, which makes bronchoscopy "a very quick, very easy, painless procedure," according to Stansell. Still, the invasive procedure is not easy to tolerate. One or both lungs are washed out (lavaged) to flush out the organisms, and brushings of the bronchial walls may also be done. The fluid and brushings are then examined by microscope. Fortunately, bronchoscopy tends to be extremely accurate at producing a confirmed diagnosis. A far more invasive approach, a surgical procedure called an open-lung biopsy, is now only very rarely used to diagnose HIV-related pneumonia.
Recently, researchers from the National Institutes of Health have advocated treatment for AIDS-related pneumonia, including PCP, without necessarily looking for the Pneumocystis organism, if cost-effective facilities for PCP testing are not available (see Research Notes, this issue).
General guidelines for treatment of all pneumonias emphasize adequate hydration to help clear secretions. If a patient is so sick that adequate intake of fluids by mouth is difficult, intravenous supplementation may be needed.
The drug therapy used depends on the organism responsible. Pneumonias caused by fungi such as Cryptococcus or Histoplasma are treated with the same antifungals, such as amphotericin B, itraconazole and fluconazole, that are used to treat other ailments caused by these organisms. Similarly, the relatively rare pneumonia caused by cytomegalovirus (CMV) is treated with the familiar anti-CMV drugs ganciclovir and foscarnet.
Treatment of bacterial pneumonias may require any of numerous antibiotics, including penicillin, amoxicillin, TMP-SMX, clarithromycin, azithromycin and the cephalosporins, depending on the specific bacteria involved. A potential source of difficulty is the increasing occurrence of drug-resistant strains, and the CDC recommends drug sensitivity testing to ensure that an appropriate therapy is chosen.
Although PCP treatment has improved steadily over time, the picture has not changed dramatically since BETA published "Guidelines for Treatment of Pneumocystis carinii Pneumonia," by editor Ronald Baker, PhD, in March, 1993. TMP-SMX remains the first-line treatment, whether administered intravenously for severe cases or orally for milder ones. The drug's toxicities may be problematic, but can often be controlled with dose adjustments and/or use of antihistamines and fever-reducing drugs. Still, in some cases it is necessary to switch to another drug.
In severe cases that require intravenous (IV) therapy, IV pentamidine remains the second-line drug of choice. Like TMP-SMX, it is extremely effective against PCP. However, IV pentamidine can produce serious, sometimes irreversible side effects, such as abnormally low blood pressure, or blood sugar level abnormalities including diabetes, pancreatitis (pancreas inflammation), kidney toxicity and loss of white blood cells. Because of these toxicities, most physicians use IV pentamidine only in the worst cases, and only after first trying TMP-SMX.
In recent years, trimetrexate has emerged as a third-line IV therapy for PCP. The drug must be administered with a second drug, leucovorin, to counter its toxic side effects. In an ACTG-sponsored study comparing trimetrexate plus leucovorin to TMP-SMX in patients with moderate to severe PCP, trimetrexate was better tolerated but less effective; the trimetrexate group had a significantly higher death rate. The high cost of trimetrexate-leucovorin is another disadvantage.
In moderate to severe cases, using a short course of corticosteroids such as prednisone along with antibiotic therapy has been shown to reduce the chances of respiratory failure and death. These regimens do not affect P. carinii directly; rather, they apparently decrease the body's inflammatory response to the organism, thereby alleviating breathing difficulties.
Several options are now available for people who cannot tolerate TMP-SMX but do not need IV therapy. The first fallback option is usually dapsone-trimethoprim or clindamycin-primaquine. Dapsone-trimethoprim, Leoung notes, is simple to administer and generally easier to tolerate than clindamycin, but some physicians prefer clindamycin-primaquine. In a recent 91-patient Canadian study, clindamycin-primaquine worked roughly as well as TMP-SMX, with fewer serious side effects but greater incidence of rash.
A more recent entry among the oral drugs is atovaquone (Mepron). Manufacturer Burroughs Wellcome (now Glaxo-Wellcome) took offense when Baker listed atovaquone as the fourth-line therapy 3 years ago (BETA, pp 18-19, March 1993). Data obtained since then, as well as a new formulation, have only somewhat boosted the drug's position. Atovaquone's strongest positive attribute is its low level of side effects. In studies comparing it to either TMP-SMX or pentamidine, atovaquone generally was less active against PCP but better tolerated--enough so that significantly more patients completed therapy without having to switch drugs.
A new, better-absorbed liquid formulation could improve atovaquone's efficacy, but the new formulation came with new labeling reducing the dosing from 3 to 2 times a day--producing approximately the same blood levels of the drug. To get new labeling, the company will need to do further research on the new formulation; in the meantime, doctors are not sure how best to use the drug. Leoung comments that "people have always had a little concern about using atovaquone alone. Most doctors who've had experience with atovaquone say, 'Well, I wouldn't give it to somebody who's really sick.' We don't know yet."
As with all areas of HIV/AIDS care, optimal prevention and treatment of pneumonia are facilitated when patients take an active role in their care. Persons who carefully note and report symptoms and side effects to their health care providers and who ask questions when they are unsure about tests or treatments improve their chances for successful care. In the present cost-conscious managed-care environment, doctors can be under tremendous pressure to do things as quickly and inexpensively as possible. "Be proactive," Leoung advises, "Speak up."
Baker RA and J Kapusik-Uner. Guidelines for treatment of Pneumocystis carinii pneumonia (PCP). BETA 18-19. March 1993.
Bozette S and others. A randomized trial of 3 anti-pneumocystis agents in patients with advanced human immunodeficiency virus infection. New England Journal of Medicine 332(11): 693-699. March 16, 1995.
Centers for Disease Control and Prevention. 1995 revised guidelines for prophylaxis against Pneumocystis carinii pneumonia for children infected with or perinatally exposed to human immunodeficiency virus. Morbidity and Mortality Weekly Report 44(RR-4). April 28, 1995.
Centers for Disease Control and Prevention. Bacterial respiratory infections, prevention of exposure. Morbidity and Mortality Weekly Report 44(RR-8). July 14, 1995.
Clumeck N. Primary prophylaxis against opportunistic infections in patients with AIDS. New England Journal of Medicine 332(11): 739-740. March 16, 1995.
Dohn M and others. Oral atovaquone compared with intravenous pentamidine for Pneumocystis carinii pneumonia in patients with AIDS. Annals of Internal Medicine 121(3):174-180. August 1, 1994.
Hirschtick RE and others. Bacterial pneumonia in persons infected with the human immunodeficiency virus. New England Journal of Medicine 333(13): 845-51. September 28, 1995.
Hughes W and others. Comparison of atovaquone (566C80) with trimethoprim-sulfamethoxazole to treat Pneumocystis carinii pneumonia in patients with AIDS. New England Journal of Medicine 328(21): 1521-1527. May 27, 1993.
Ioannidis J and others. A meta-analysis of the relative efficacy and toxicity of Pneumocystis carinii prophylactic regimens. Archives of Internal Medicine. 156: l77-188. Jan. 22, 1996.
Luce J and MJ Clement. Pulmonary diagnostic evaluation in patients suspected of having an HIV-related disease. Seminars in Respiratory Infections. 4(2): 93-101. June 1989.
Moe AA and WD Hardy. Pneumocystis carinii infection in the HIV-positive patient. Journal of the Physicians Association for AIDS Care 20-35. June 1994.
Rubin RH. Infection in the immunosuppressed host. Scientific American Medicine 7:X:1-23. July, 1994.
Rubin RH and R Greene. Clinical approach to the compromised host with fever and pulmonary infiltrates. Clinical Approach to Infection in the Compromised Host. Plenum Medical Book Company, New York, NY, 1994.
Sattler FR and others. Trimetrexate with leucovorin versus trimethoprim-sulfamethoxazole for moderate to severe episodes of Pneumocystis carinii pneumonia in patients with AIDS: a prospective, controlled multicenter investigation of the AIDS Clinical Trials Group protocol 029/031. Journal of Infectious Diseases 170: 165-172. July 1994.
Schwab D. Sulfa desensitization. Project Inform Fact Sheet Supplement. May 23, 1993. Selik RM and others. Trends in infectious diseases and cancers among persons dying from HIV infection, USA, 1987-1992. 35th Interscience Conference on Antimicrobial Agents and Chemotherapy. September 17-20, 1995. Abstract #I-22.
Toma E and others. Clindamycin/primaquine (C/P) vs. trimethoprim-sulfamethoxazole (T/S) for PCP in AIDS. 35th Interscience Conference on Antimicrobial Agents and Chemotherapy. September 17-20, 1995. Abstract #LM96.