Stopping PCP prophylaxis?

Zellweger (2004) studied the incidence of PCP in 78 HIV-positive patients on HAART who stopped secondary prophylaxis with a CD4 cell count >200 cells/mm³ and CD4 percentage of 14%. Median CD4 cell count was 380 cells/mm³ when prophylaxis was stopped. In a median follow-up of 40.2 months, there were no cases of recurrent PCP.

Lopez Bernaldo de Quiros (2000) studied 474 people on antiretroviral therapy with CD4 counts above 200 for at least 3 months. The patients were all taking primary PCP prophylaxis and were randomised to cease or continue prophylaxis. The group had a median CD4 count of 342 and 38% had viral load above 500 copies although all were below 5000 copies. 121/474 participants had a previous CD4 count below 50 cells/mm3. At 20 months, no cases of PCP occurred among the group who ceased prophylaxis. 113 people on secondary prophylaxis after PCP diagnosis were also randomised to cease or continue prophylaxis after they had sustained a CD4 count above 200 for 3 months. 68% had had a previous CD4 count below 50. Over two years had passed since their PCP episode in 54% of cases. Median CD4 count was 355, 76% had undetectable viral load and 24% were between 500-5000 copies. At 12 months, no cases of PCP occurred among the 60 people who stopped prophylaxis.

Ledergerber (2001) studied 325 people on HAART from 8 prospective European cohort studies who ceased their secondary PCP prophylaxis between 1996-2000. All had CD4 counts above 200 when they stopped treatment and the median CD4 count was 350. Average lowest ever CD4 count was 50 in this group. During 13 months follow-up after cessation of PCP prophylaxis, no cases of PCP occurred. The rate of bacterial infections during follow-up was low at 2.7 per 100 person-years.

Mussini (2000) randomised 708 people to either continue or discontinue PCP prophylaxis after sustaining a CD4 count above 200 for at least 3 months. After an average of 6.4 months follow-up, no cases of PCP occurred. No cases of toxoplasmosis were reported among the 233 people who were also at risk of that OI (those with previously active toxoplasmosis were excluded from the study).

Furrer (2001) conducted a prospective study of 262 people who discontinued PCP prophylaxis after their CD4 counts increased to above 200 for at least 12 weeks. Median CD4 at discontinuation was 325 (range 210-806) and the median CD4 nadir prior to successful anti-HIV treatment was 110 (range 0-240). Average follow-up time was 11.3 months (range 3-18.8 months). No cases of PCP or toxoplasmosis were reported; 9 people recommenced prophylaxis and 2 subsequently died. No cases of PCP occurred among those at greater risk of PCP (low CD4 nadirs or detectable viral load). One case occurred in a person with a CD4 count of 530 and a viral load below 50 copies.

Valentine (2000) reported on a person who developed PCP despite a CD4 count over 400 after immune recovery due to HAART. No PCP antigen-specific immune response was detected.

Kirk (1999) reported that PCP prophylaxis was stopped among 193 people on primary prophylaxis and 26 people on secondary prophylaxis after their CD4 counts rose above 200, or they had been on HAART for at least 6 months. One case of PCP was reported.

Weverling (1999) reported observational data from the EuroSIDA cohort. 378 people stopped PCP prophylaxis after an average of ten months on highly active antiretroviral therapy (HAART). 319 of these had never had a bout of PCP before. The average lowest CD4 count in this sub-group was 123 cells, compared with 60 cells in the group of 59 people who were on maintenance therapy for PCP. Prophylaxis was stopped when the CD4 count had risen, on average, to 270 cells (the same for both groups), and had been taken for an average of 16 months. CD4 counts had been sustained above 200 for an average of 8 months before prophylaxis was stopped in the primary prophylaxis group, and 5 months in the secondary prophylaxis group. Over a follow-up period of 7 months off prophylaxis in the primary group; 5 in the secondary, no cases of PCP were observed. The authors suggested that the risk posed by stopping PCP prophylaxis following a sustained increase in CD4 above 200 is sufficiently low in people who have never had PCP to warrant discontinuation. However, they advocate further study before such guidance can be offered to people taking secondary prophylaxis.

Schneider (1999) reported data from a prospective observational study of 78 people on HAART who stopped PCP prophylaxis as soon as they had two CD4 counts (at least one month apart) above 200 cells. 62 people were receiving primary prophylaxis and the other 16 secondary prophylaxis. During prophylaxis, the lowest mean CD4 count had been 79 cells. At the time of stopping prophylaxis, mean CD4 was 347 cells, and patients had been on HAART for a median of 9.8 months. Viral load was undetectable at the time of stopping in 61 patients. The group were followed for an average of 12.7 months since discontinuation, and there were no cases of PCP in this time.

Dworkin (2001) reported on 626 people (group 1, previous CD4 count below 200, CD4 count rise of over 100 on treatment, subsequent CD4 count of over 200) and 3497 people (group 2, never had a CD4 count below 200). Group 1 had either never started or had stopped PCP prophylaxis. Analysis found no statistical difference in the incidence of PCP between the two groups. 3 cases of PCP occurred over 471 person years (0,8 per 100 person years) and 32 occurred over 4200 person years (0.7 per 100 person years). One case of PCP occurred among the group with a CD4 count previously below 25, at a rate of 4.4 cases per 100 person years. This was not statistically significant but researchers suggested that long-term follow-up may reveal that those with the most advanced HIV disease are most at risk of PCP if they stop their prophylaxis.

Koletar (2001) (ACTG 888) investigated discontinuation of PCP prophylaxis in 144 people who had had a previous CD4 count below 100 (group 1) and 129 people who had had previous PCP in the 6 months prior to entering the study (group II). CD4 levels were above 200 when prophylaxis was discontinued. After median follow-up times of 82 weeks in group I, and 63 weeks in group II, no cases of PCP occurred. 8 people resumed prophylaxis when their CD4 counts fell below 200.

Yangco (2000) conducted a non-randomised, prospective study of 146 patients who ceased PCP prophylaxis and 345 who continued prophylaxis. All had sustained CD4 count rises over 200. Those who stopped were followed for a mean follow-up time of 18.2 months whereas those who continued were followed for 14 months. No cases of PCP nor toxoplasmosis were reported in either group. The group who continued prophylaxis had more advanced disease as indicated by lower mean CD4 nadir and higher viral load.

Co-trimoxazole as PCP prophylaxis for adults<3>

Co-trimoxazole is the treatment of choice for the prevention of PCP. It has also been evaluated for the prevention of other bacterial infections and malaria among people with HIV in Africa, where PCP is rarely diagnosed among adults.The African research agenda has included the need to assess patterns of resistance to cotrimoxazole among pathogens contributing to the illness burden for people with HIV. There have also been efforts to define sub-groups of people with HIV - for example, those co-infected with TB - who could derive maximum benefit from co-trimoxaxole prophylaxis. The level of benefit also depends strongly on the extent to which people adhere to the treatment - which may be more important in the case of toxoplasmosis than it is for preventing PCP or other conditions. These questions will almost certainly have different answers for different populations at different times.

The US Public Health Service Task Force on PCP prophylaxis recommended co-trimoxazole (one double-strength tablet every day) for both primary and secondary prophylaxis of PCP in HIV-infected people with no history of severe reactions to co-trimoxazole and either prior PCP, a CD4 count below 200 or unexplained fever or thrush (CDC, 1992).

Dworkin (1998) from the CDC reported on more than 19,000 US HIV-infected patients analysed in terms of prophylatic treatment with cotrimoxazole. Prophylactic treatment was associated with reduced PCP, hemophilus, salmonella, toxoplasma and staphylococcus infections.

Anglaret (1999) enrolled 545 people in Abidjan, Cote d'Ivoire, with stage 2 or 3 HIV disease who received co-trimoxazole or placebo. One third of participants were followed for 12 months. Data showed that significantly fewer patients in the co-trimoxazole group experienced at least one severe event compared to placebo (84 vs 124). There was 120 events in the treatment group, compared with 198 in the placebo group. The treatment group were much more likely to remain free from any severe event. The diseases that occurred less often in the treatment group were: bacterial pneumonia, malaria, isosporosis, and acute unexplained fever. Overall survival was not affected by co-trimoxazole (but see below).

Losina (2003) compared deaths that were not attributed to diagnosed OIs among people with HIV at different CD4 counts, in the placebo arm of the Anglaret study, with and without a history of OIs. This 'chronic mortality' was markedly higher among people who did have such a history, which was taken as strengthening the case for prophylaxis. This study implies that although not diagnosed with OIs, the deaths were probably due in many cases to treatable or preventable infections.

Leoung (2001) conducted a randomised, double-blind trial comparing 2 methods of co-trimoxazole prophylaxis reintroduction in HIV-infected people who had had an allergic reaction to co-trimoxazole: a 6-day dose escalation versus direct rechallenge. 75% of the dose-escalation group and 57% of the direct-rechallenge group took daily single-strength co-trimoxazole for 6 months (p=0.014). Cessation of prophylaxis due to adverse reactions was more common among the direct-rechallenge group.

Jourdain (2000) studied a cohort of 206 pregnant women, 64% of whom were treated with co-trimoxazole. The average CD4 count in the women was 120. The incidence of PCP was 0% in the treated group versus 13% in the untreated group. Babies born to untreated women were more likely to be born before 36 weeks and/or at a birth weight below 2500gms. The authors concluded cotrimoxazole is safe for use during pregnancy.

Saah (1995) reported that the most significant risk factor for failure of PCP prophylaxis is a CD4 count below 75. Among 476 people in the MACS cohort receiving PCP prophylaxis, 20% developed PCP. More than three-quarters of these had CD4 counts below 50, and 85% had counts below 75.

Fischl (1988) conducted a placebo-controlled trial of co-trimoxazole (160mg TMP plus 800mg SMX twice a day) as primary PCP prophylaxis in 60 people with KS and no history of opportunistic infections. After three years, 53% of the placebo group had developed PCP and 93% of the group died; none of the treated group developed PCP and only 60% died.

Several studies have shown that co-trimoxazole (1 double strength tablet) delivered three times a week (as opposed to daily) is effective as primary prophylaxis for PCP with reduced incidence of adverse reactions. Ruskin (1991) conducted a retrospective analysis of co-trimoxazole three times weekly as PCP prophylaxis in 116 patients (71 had a prior history of active PCP, mean follow-up 18.5 months; 45 never had PCP, but CD4 below 200; mean follow-up 24.2 months). 0/116 developed PCP. 28% had side effects (rash, pruritus, nausea); 15/116 discontinued treatment, with 11/16 clearly drug intolerant.

El-Sadr (1999) enrolled 2625 people in a a randomized, prospective study comparing daily vs. thrice weekly co-trimoxazole as primary or secondary PCP prophylaxis. The overall rates of PCP were 3.5% (90 cases) and 4.1% (105 cases) in the daily and thrice-weekly groups respectively. Because of trends favouring daily co-trimoxazole for all major endpoints, the authors recommend daily co-trimoxazole as the preferred dosing.

Para (2000) reported the results of ACTG 268, indicating that initiation of PCP prophylaxis using gradually increasing doses of liquid co-trimoxazole over the first two weeks, rather than starting on double-strength tablets, was associated with a significantly lower rate of discontinuations during the first 12 weeks of treatment (17% versus 33%).

Wiktor (1999) enrolled 771 people in Cote d'Ivoire infected with HIV-1 or HIV-1 and HIV-2 who tested positive to a sputum test for TB. Participants were randomised to either one tablet of co-trimoxazole (800mg trimethoprim and 160 mg sulphamethoxazole) daily or placebo one month after starting 6-month TB treatment. Median age on entry was 32 and median CD4 count was 317. Co-trimoxazole significantly reduced death and hospital admissions. 51 people died in the co-trimoxazole group compared with 86 in the placebo group. 29 hospital admissions occurred in the treatment group, compared with 47 in the placebo group. Median follow-up was about 10 months in both groups. Hospital admissions were significantly lower for enteritis due to isosporiasis, non-typhoid salmonellas and septicaemia among the co-trimoxazole group.

Kancheya (2003) evaluated the incidence of severe bacterial infections among symptomatic people with HIV in Lusaka, Zambia. While virtually all were eligible for CTX prophylaxis under current guidelines, it was not considered practical to provide it to all of them. A scoring system including at least two of the following variables was used to identify a sub-group at greatly increased risk, who could be prioritised for CTX treatment. The variables were: prior treatment for chronic cough, chronic fever, asthenia, or night sweats, and ESR >90th percentile, haematocrit <5th percentile and BMI <10th percentile. A scoring scheme including at least two of these variables had a sensitivity of 37% and a positive predictive value of 27%.

Contrasting pictures of patterns of opportunistic infections were reported at an international conference in Paris from Haiti and Malaysia. In Malaysia (Kamarulzaman), in a predominantly male population of people with HIV, mostly acquired heterosexually or through injecting drug use, the most common opportunistic infections were tuberculosis (40%), Pneumocystis pneumonia (24%) and toxoplasmosis (14%). In Haiti, in contrast, TB accounted for between a third and a half of all presenting diagnoses, PCP for less than one per cent and toxoplasmosis was not reported (Koenig).

For research into resistance to co-trimoxazole, see the entry Pneumocystis pneumonia (PCP) - treatment key research in Symptoms and illnesses: A to Z of illnesses.

Aerosolised pentamidine as PCP prophylaxis

Hirschel (1991) conducted a placebo-controlled trial of aerosolised pentamidine (AP) (300 mg every 28 days administered by the Respirgard II nebuliser) as primary PCP prophylaxis in 223 people with CD4 counts below 200, advanced ARC or AIDS without prior PCP. Over one year, 23 cases of PCP occurred in the placebo group compared with only 8 in the treated group.

Leoung (1990) enrolled 408 participants in a dose comparison trial of aerosolised pentamidine (30 mg every two weeks, 150 mg every two weeks, or 300 mg every four weeks). Eighteen months after randomization, the participants in the 300 mg arm had had 8 confirmed episodes of PCP while receiving treatment, as compared with 22 in the 30-mg arm. The 150 mg arm had intermediate results but ones not significantly different from those of the 300 mg arm. Participants with previous episodes of PCP and CD4-cell counts less than 200 per cubic millimetre were at the highest risk for PCP.

Rizzardi (1995) reported that in a study comparing once-monthly versus twice-monthly aerosolised pentamidine, after a median observation period of 232 days, 14 relapses of PCP occurred in the once-monthly group, compared with 5 relapses in the twice-monthly group.

Several studies have also established the efficacy of aerosolised pentamidine delivered through the Fisons hand-held nebuliser (Fisoneb). Montaner (1991) conducted a placebo-controlled trial of aerosolised pentamidine as secondary PCP prophylaxis in 164 people who had had one episode of PCP and no other AIDS-defining opportunistic infection. Over a mean follow-up of 15.5 weeks, 27/32 (82%) PCP recurrences occurred in the placebo group. Murphy (1991) enrolled 175 patients in a blinded dose-comparison study of nebulised pentamidine 5 mg, 60 mg, or 120 mg every two weeks administered through the Fisons nebuliser. While the low-dose regimen was ineffective, the 60 mg and 120 mg regimens were equally effective. At 52 weeks, probability of being PCP-free was 88% in the 60 mg group and 93% in the 120 mg group. While these results establish the efficacy of nebulised pentamidine delivered through the Fisons nebuliser, a study directly comparing the Fisons and Respirgard II nebulisers has not been conducted.

Schneider (1992) randomized 215 HIV-positive patients with CD4 counts below 200 and no history of PCP to receive co-trimoxazole (one single strength or double strength tablet daily) or aerosolized pentamidine (300mg every 4 weeks via the Respirgard II nebulizer). After a mean follow-up of 264 days, 6/71 pentamidine recipients had a confirmed episode of PCP, while 0 of 412 patients taking co-trimoxazole group had PCP. Adverse reactions requiring treatment discontinuation were higher in both co-trimoxazole groups (17/71 on 480 mg/day and 18/71 on 960 mg/day) compared to the aerosolized pentamidine group (2/71).

Rizzardi (1996) randomized 220 patients to receive co-trimoxazole (one double-strength tablet daily) or aerosolized pentamidine (300mg monthly). At one year, there was a non-significant trend favouring co-trimoxazole. After two years follow-up, intent-to-treat analysis showed that 6 and 5 patients developed PCP and 30 and 18 patients died in the co-trimoxazole and pentamidine groups respectively (differences not significant). The relative risk of mortality at two years was 1.5 in the pentamidine group compared with the co-trimoxazole group.

Hardy (1992) randomised 310 AIDS patients who had recovered from one bout of PCP to receive either one double-strength co-trimoxazole tablet a day or aerosolized pentamidine 300 mg every 4 weeks (ACTG 021). After a mean follow-up of 17.4 months, PCP had recurred in 14/154 (9%) patients randomized to co-trimoxazole and 36/156 (23%) patients randomized to aerosolized pentamidine. The one-year estimated recurrence rate for the aerosol pentamidine group was 18.5%, compared with 3.5% for the co-trimoxazole group. The risk of recurrence in the pentamidine group was 3.25 times that in the co-trimoxazole group. Adverse reactions requiring a change in treatment occurred in 42/154 (27%) of co-trimoxazole recipients and 6/156 (4%) of aerosolized pentamidine recipients. Rash and fever accounted for most of the toxicities. There were no differences in the rates of hæ­¡tologic or hepatic abnormalities between the two groups. Survival was equivalent in the two groups.

A cost-benefit analysis performed by Castellano showed that first-line co-trimoxazole (1 double-strength tablet/day) resulted in better survival and lower cost than first-line AP (300 mg/month). This literature review determined that median survival after initiating secondary prophylaxis with co-trimoxazole was 2.050 years, compared with 1.983 years with AP. (The median survival without any prophylaxis was 1.345 years.)

Dapsone as PCP prophylaxis

Blum (1992) prospectively compared dapsone (100mg/day) and co-trimoxazole (one double-strength tablet/day) as primary prophylaxis for PCP in HIV-positive patients with CD4 counts below 200. After 1,638 patient months of follow-up, 1/47 patients randomized to dapsone and 1/39 patients randomized to co-trimoxazole developed PCP. 33/47 dapsone recipients and 25/39 co-trimoxazole recipients discontinued their study therapy, most frequently for rash. Ten patients switched from dapsone to co-trimoxazole, 4 successfully; and 11 patients crossed over from co-trimoxazole to dapsone, 6 successfully.

Slavin (1992) conducted a randomized comparison of dapsone (100mg PO twice weekly) and aerosolized pentamidine (400mg nebulized monthly) for PCP prophylaxis in patients with CD4 counts below 200 or a previous episode of PCP. 37/50 dapsone recipients and 32/46 AP recipients had had a previous episode of PCP. Mean CD4 counts at entry were 128 and 124 in the dapsone and AP groups, respectively. All patients received AZT. After a median follow-up of 18 months, 9/50 dapsone recipients and 8/46 AP recipients had developed PCP (no difference). The rates of adverse reactions were equal in the two groups (4 rash and 2 nausea on dapsone, and 5 cough and 1 dizziness on AP).

Bozzette (1995) enrolled 843 people with CD4 counts below 200 and no prior history of PCP in a trial (ACTG 081) comparing three regimens for the primary prophylaxis of PCP. Participants were randomized to receive co-trimoxazole (one double-strength tablet daily), aerosol pentamidine (300mg monthly), or dapsone (50mg twice daily). The median follow-up was 39 months. Upon intent-to-treat analysis, PCP occurred with equal frequency in the three groups. The estimated 36-month cumulative risks of developing PCP were 18%, 17% and 21% for the co-trimoxazole, dapsone and aerosol pentamidine groups, respectively. The difference in risk was negligible for participants entering the study with CD4 counts of 100 or above; however, for those with baseline CD4 counts below 100 the risk was 33% for aerosolised pentamidine, 22% for dapsone and 19% for co-trimoxazole. Survival was equivalent in the three groups (134, 131, and 138 deaths). Significantly fewer side effects requiring a switch in treatment assignment occurred in the aerosol pentamidine group. The mean time to discontinuation of assigned treatment was 14.6, 17.1, and 13.7 months in the co-trimoxazole, aerosol pentamidine, and dapsone groups respectively.

Martin (1992) conducted a retrospective chart review of 211 HIV-infected patients, 131 of whom (62%) received primary prophylaxis and 80 of (38%) whom received secondary prophylaxis for PCP. First-line treatment was co-trimoxazole (1 double-strength tablet twice a day, thrice weekly; n=133; mean duration 7.4 months); second-line was dapsone (50mg/day; n=77; mean duration 5.7 months); third choice was aerosolised pentamidine (300mg once monthly; n=125; mean duration 9.3 months). 73% received concomitant antiretroviral therapy. PCP did not develop in those receiving co-trimoxazole, compared with 5 cases on dapsone and 17 on aerosolised pentamidine. 56% of the co-trimoxazole group and 55% of the dapsone group changed drug due to adverse reactions; only 2% in the aerosolised pentamidine group required drug change.

Mallolas (1993) randomized 331 HIV-infected patients with CD4 counts below 200 or a previous diagnosis of AIDS to receive dapsone/pyrimethamine (100/25mg weekly), aerosolized pentamidine (300mg monthly) or co-trimoxazole (160/800mg thrice weekly). The mean follow-up was 313 days. Upon intention-to-treat analysis, the PCP rates per year of observation were 8.3%, 5.6%, and 3.0% in the D/P, AP and co-trimoxazole groups, respectively. These differences were not statistically significant.

Opravil (1995) randomized 500 patients to receive either dapsone/pyrimethamine (D/P) (200mg/75mg once weekly) or aerosolized pentamidine (300mg once monthly via Respirgard II nebulizer). All patients had CD4 counts below 200 at baseline and 20% had had previous PCP. After more than 9 months' mean follow-up, intent-to-treat analysis showed that 7/273 dapsone/pyrimethamine recipients and 9/227 AP recipients had developed PCP (no significant difference). Intolerance requiring a switch in treatment occurred in 36% of dapsone/pyrimethamine recipients and 4% of AP recipients. Most frequent side effects were nausea, fever, and hæ­¡tologic toxicity.

Girard (1993) randomized 349 symptomatic patients with CD4 counts below 200 to receive dapsone/pyrimethamine (50mg daily/50mg weekly) or aerosolized pentamidine (300mg monthly) on an open-label basis. After a median follow-up of 539 days, 10/173 dapsone/pyrimethamine recipients and 10/176 aerosolized pentamidine recipients had developed PCP. 42 dapsone/pyrimethamine recipients and 3 aerosolized pentamidine recipients discontinued therapy due to toxicity. No difference in survival was observed.

Salmon-Ceron (1995) randomised 200 people with recently treated PCP to receive either monthly aerosolised pentamidine or daily dapsone (50 mg) plus iron protoxolate. The trial was halted early after an average follow-up of 12-13 months, when 42% of participants assigned to dapsone had died compared with 21% assigned to pentamidine.

Antinori (1992) randomized 197 patients with CD4 counts below 200 and no prior PCP to receive dapsone/pyrimethamine (100mg weekly/25mg twice weekly), aerosolized pentamidine (300 mg monthly via Respigard II nebulizer), or co-trimoxazole (one double-strength tablet every other day). After a median follow-up of 7.7 months, PCP had developed in 9/63, 4/68 and 1/66 of dapsone/pyrimethamine, aerosolized pentamidine and co-trimoxazole recipients respectively. Upon intent-to-treat analysis, the difference between the co-trimoxazole and D/P groups was significant. Other comparisons did not show significant differences.

Podzamczer (1993) randomized 166 people with CD4 counts below 200 to receive dapsone/pyrimethamine (DP 100mg/50mg once weekly) or co-trimoxazole (2 double-strength tablets three times weekly). After a median follow-up of 380 days, intent-to-treat analysis found that PCP occurred in 15.2% of DP recipients and 3.7% of the co-trimoxazole recipients (p=0.01). At 12 and 24 months, the cumulative rates of PCP were 5% and 42% among the DP group and 3% and 10% among the co-trimoxazole group. 67% of the co-trimoxazole group and 42% of the DP group experienced side effects and the drop out rate due to toxicity was 12% vs 2% (p=0.01).

Podzamczer (1995) randomised 230 HIV-infected people with CD4 counts below 200 to trimethoprim-sulfamethoxazole (160mg/800mg twice a day, thrice weekly) or dapsone/pyrimethamine (DP 100mg/50mg twice weekly). After a median follow-up of 430 days, 6/96 (6.3%) people taking DP and 0 of 104 people taking co-trimoxazole developed PCP(p<0.0001). Only one episode of PCP developed while patients were taking prophylaxis.

Other prophylactic regimens

Chan (1999) compared two atovaquone suspension regimens (750mg and 1500mg once daily) with aerosolized pentamidine (300mg once a month) for the prevention of PCP in people with HIV who were intolerant of trimethoprim or sulfonamides. Of 549 people enrolled, 42% had had a prior episode of PCP. Median CD4 cell count at baseline was 57. 73% were receiving anti-HIV therapy at entry; 28% monotherapy. By the end of the study, 48% had received a PI. Median time on therapy was 6.6 months and median follow-up was 11.3 months. An intent to treat analysis found no significant differences in proportions developing PCP (26%, 22%, 17% respectively), or mortality (20%, 13%, 18% respectively). Incidence of PCP in secondary prophylaxis recipients was more than twice that in the primary group (33% vs. 13%), though there were no differences across treatment arms. Similarly, there were more deaths amongst those on secondary prophylaxis (27% vs. 13%). Half of the cohort were receiving aerosolized pentamidine at baseline, but this made no difference to the overall outcome in terms of the incidence of PCP or AEs. There were more adverse events (rash, fever, vomiting and anaemia) amongst those receiving atovaquone. There was no apparent relationship between dose of atovaquone and incidence of AEs.

El-Sadr (1998) compared atovaquone suspension (1500mg daily) versus dapsone (100mg daily) among 1057 people who had CD4 counts below 100, and were intolerant to co-trimoxazole. Median time on dapsone was 10.9 months for those on dapsone at baseline and 3.9 months not previously taking dapsone. Time on atovaquone was about seven months, and was not affected by previous use of atovaquone.There was no difference in the rates of development of PCP or toxoplasmosis between the two groups although the low number of cases of toxoplasmosis meant the researchers would not draw conclusions about the drugs' comparative ability to prevent this infection. Among participants who were not receiving dapsone on entry, atovaquone was much better tolerated than dapsone. Among participants who were already receiving dapsone at study entry, dapsone was better tolerated. Hence, atovaquone may be the preferred prophylaxis among people intolerant to co-trimoxazole who have not commmenced dapsone prophylaxis. Researchers could not explain why those who discontinued treatment had lower rates of PCP compared with those who remained on treatment.

Correspondence subsequent to the publication of El-Sadr's study suggested that dapsone and atovaquone have equivalent efficacy in preventing PCP and that the much greater cost of atovaquone means that dapsone should be the first choice in cases of intolerance to co-trimoxazole (Horowitz 1999).

Dunne (1999) reported on occurrence of PCP in a study of MAI prophylaxis comparing azithromycin, rifabutin or the combination of both drugs for a mean duration of approximately one year. 97% of participants were receiving specific PCP prophylaxis. The incidence of PCP was 8.7% among people on azithromycin containing regimens, compared with 13% among those on rifabutin. Thus, the use of azithromycin in concert with other PCP prophylactic regimens reduces the risk of PCP by one half.

Cheung (1993) conducted a retrospective review of 96 patients who received intramuscular pentamidine 300mg or 4mg/kg monthly as prophylaxis for PCP. 49/96 had prior PCP, the remainder had CD4 count below 200. After 7-8 months follow-up, 3 patients developed PCP. Toxicities observed were 2 cases of hypotension, three sterile abscesses and three cases of glucose intolerance.

Smoking and PCP

Wewers (1998) reported that smoking is associated with a weakened immune system in the lungs.CD4 and CD8 tests conducted on bronchoalveloar lavage fluid found that smokers had lower counts than non-smokers. Cytokine production was also reduced among the smokers.

Nieman (1993) reported on a case series of 84 individuals with AIDS who provided accurate details of their smoking habits before their AIDS-defining diagnosis. Progression time to AIDS (all diagnoses) was significantly reduced in HIV-1-seropositive smokers: median time to AIDS was 8.17 months for smokers (n = 43) and 14.50 months for non-smokers (n = 41). Smokers developed Pneumocystis carinii pneumonia (PCP) more rapidly than non-smokers, with a median time to PCP of 9.0 months, compared with 16.0 months for non-smokers. Smoking had no significant effect on progression time to AIDS when not due to PCP.

Buskin (1992) examined the medical records of 1204 HIV seropositive persons at seven outpatient sites in Seattle during 1990-1991. To exclude persons at minimal risk of PCP, the analysis was limited to persons with CD4 counts less than 500 with known smoking status (n = 598). There were 325 current smokers and 273 nonsmokers. Of the nonsmokers, 54 were known to be ex-smokers, 29 were never smokers, and the remaining 190 (70%) could not be further classified. Because the number of never smokers was small and because many current nonsmokers were likely to be ex-smokers with a high risk of PCP (odds ratio = 5.2 for ex-smokers relative to light smokers), light smokers (lowest pack-year tercile, less than 12 pack-years) were used as a reference group. One hundred and twenty-six persons (21%) had one or more episode(s) of PCP. There was a strong dose-response effect for intermediate smokers (12 to 19 pack-years) and heavy smokers (20+ pack-years) when looking at the outcome of ever having had PCP. The crude relative odds of PCP for the intermediate and heavy smokers were 3.1 and 5.6, respectively. After adjustment for CD4 count, gender, ethnicity, and use of PCP prophylaxis by logistic regression, the odds of PCP were 2.9 for intermediate and 3.5 for heavy smokers. Heavy smokers were over 3 times more likely to have had PCP than were light smokers.

Other risk factors

Navin (2000) reviewed 211 cases of pneumonia among HIV-infected individuals and compared results to matched HIV-infected individuals with other infections. Multivariate analysis found the following factors associated with PCP: breathing chemical irritants such as insect spray, petrol or paint fumes in the previous 6 months; hospitalisation for pneumonia in the last 6 months; camped or hiked in the last 6 months; and, gardening in the last 6 months.

PCP prophylaxis for children

References

See PCP - treatment research and references for references to studies cited above.