Outcome from treatment of Pneumocystis jirovecii pneumonia with co-trimoxazole

Abstract

A retrospective case-notes audit of 359 HIV-1-infected adult patients with first-episode laboratory-confirmed Pneumocystis jirovecii pneumonia treated with co-trimoxazole (from 1987 adjuvant steroids were used if PaO₂ <9.3 kPa) showed that only 230/359 (64%) patients completed treatment; 104 (29%) patients had treatment-limiting toxicity; rash occurred in 4/60 (6.7%) patients in 1985–1988 and in 15/47 (31.9%) in 2005–2008. Twenty-five patients (7%) failed co-trimoxazole treatment. Overall mortality was 13.6% (49/359); mortality among patients who failed co-trimoxazole treatment was 48% (12/25) and by contrast mortality was 4.8% (5/104) among patients with treatment-limiting toxicity.

Keywords

co-trimoxazole toxicity treatment failure

INTRODUCTION

Early reports of the treatment of HIV-associated Pneumocystis jirovecii pneumonia noted high rates of treatment-limiting adverse drug reactions (ADRs) and treatment failure with both co-trimoxazole and intravenous pentamidine.^1–3 In one of these studies, co-trimoxazole was associated with fewer treatment-limiting ADRs and improved survival compared with intravenous pentamidine.² Since then co-trimoxazole has been the preferred treatment for P. jirovecii pneumonia based on its perceived efficacy, a faster clinical response than intravenous pentamidine² and its comparable intravenous and oral bioavailability. However, treatment-limiting ADRs are common.^1–6 We recently noted that several patients being treated with co-trimoxazole for Pneumocystis pneumonia (PCP) developed treatment-limiting rash, prompting us to audit outcome from the treatment of HIV-associated PCP using co-trimoxazole.

METHODS

A retrospective case-notes audit of 359 HIV-1-infected adult patients admitted to the HIV/AIDS inpatient ward at University College London Hospitals (formerly The Middlesex Hospital) was conducted. All patients had first-episode laboratory-confirmed PCP treated with co-trimoxazole; adjuvant steroids were used (from 1987) if PaO₂ <9.3 kPa. Outcome was defined as completed (21 days) treatment, or a regimen change due to treatment-limiting toxicity or failure. Toxicity was recorded as follows: dermatological = severe rash, not present prior to starting co-trimoxazole; haematological = fall in neutrophil count or haemoglobin; renal = rise in serum creatinine to a >× 2 upper limit of normal; gastrointestinal = severe nausea or vomiting for ≥2 days, or rise in alanine aminotransferase enzyme levels to a >× 5 upper limit of normal; drug fever = new onset of fever after the patient had been afebrile on therapy. Treatment failure was defined as persistent fever and worsening hypoxia, and/or radiographic deterioration, occurring after a minimum of five days of co-trimoxazole.⁷ Survival was defined as being alive one month after completing not a must therapy.

RESULTS

The 359 patients (355 men) were aged 37 years (31–42) {median (interquartile range [IQR])}; their HIV risk factor was men who have sex with men = 292 (81.3%), heterosexual sex = 54 (15%), injection drug use = 3 (0.8%) and unknown = 10 (2.8%). For 328 patients (91.4%) PCP was AIDS defining. The median (IQR) CD4 count was 45 cells/µL (20–100) and admission PaO₂ (breathing room air) = 8.8 kPa (7.5–10.2) (median [IQR]); 190 (53%) patients received adjunctive steroids. Fifty-three patients (14.8%) required intensive care unit (ICU) admission, of whom 23 were mechanically ventilated. Twenty-nine patients developed a pneumothorax (nine were mechanically ventilated). Overall mortality = 13.6% (49/359); before June 1996 mortality = 16.9% (35/207) and after June 1996 (when highly active antiretroviral therapy [HAART] was first available) mortality = 9.2% (14/152). No patient was receiving HAART at diagnosis of PCP.

Only 230/359 (64%) patients completed first-line co-trimoxazole treatment; 104 (29%) patients had treatment-limiting toxicity; dermatological (rash) = 63, haematological = 15, gastrointestinal (hepatitis = 11, vomiting = 9), drug fever = 3 and renal failure = 3. Rash occurred in 4/60 (6.7%) patients in 1985–1988 and in 15/47 (31.9%) in 2005–2008; chi square (for trend) = 18.56, P = 0.001 (Table 1).

Table 1

Frequency of treatment-limiting rash by time period among 359 patients with laboratory-confirmed PCP who were treated with co-trimoxazole as first-line therapy

Time period (years)	No. with rash/No. with PCP	%
1985–1988	4/60	6.7
1989–1992	6/63	9.5
1993–1996	24/99	24.2
1997–2000	11/41	22.9
2001–2004	5/42	11.9
2005–2008	15/47	31.9
Total	63/359	17.5

PCP = Pneumocystis pneumonia

Twenty-five patients (7%) failed co-trimoxazole treatment. At presentation, patients subsequently failing treatment were sicker (median PaO₂ = 6.8 kPa) than those completing or experiencing treatment-limiting toxicity (median = 9.1 and 9.1 kPa, respectively; P < 0.001). Of those failing co-trimoxazole, 17 (68%) needed ICU admission and six (24%) developed a pneumothorax. Mortality among those failing co-trimoxazole = (12/25) = 48%; by contrast mortality = (5/104) = 4.8% among patients with treatment-limiting toxicity; P < 0.001.

Patients failing co-trimoxazole or with treatment-limiting toxicity were increasingly likely (with time) to receive clindamycin/primaquine as second-line therapy; chi square (for trend) = 11.26 and 53.6; P = 0.004 and 0.031, respectively.

DISCUSSION

This retrospective single-centre audit demonstrates that co-trimoxazole is effective treatment for PCP in HIV-infected adults. However, only 64% of patients successfully completed this treatment; 28% and 7% changed therapy because of toxicity and failure, respectively. We audited only patients with first-episode PCP and excluded patients with second- and third-episode PCP as the mortality rate in these patients is higher than in patients with first-episode PCP⁸; these patients may also have previously experienced co-trimoxazole toxicity or treatment failure in their first episode.

The striking observation that rash was commoner in more recent years may be due to several factors. First, more recently, clinicians may have had increasing confidence in the efficacy/lack of toxicity of second-line therapy using clindamycin with primaquine, whereas in earlier years they may have ‘treated-through’ because of concerns about the lack of efficacy/toxicity of intravenous pentamidine.⁷ Second, the toxicity of sulphamethoxazole is ascribed to its metabolism via hepatic cytochrome P450, to a hydroxylamine moiety. Co-administration of fluconazole inhibits the oxidation of sulphamethoxazole to hydroxylamine.⁹ In recent years clinicians less often prescribed fluconazole therapy to patients with PCP. The high rate of observed rash among patients with PCP cautions against early institution of HAART with agents known to cause rash, as ascribing a cause may be problematic.¹⁰

Treatment-limiting toxicity leading to a change to second-line therapy was not associated with a poor outcome; by contrast, mortality was high among those failing to respond to co-trimoxazole. These observations confirm those of Klein et al. ³, who reported mortality rates of 46% and 3% among patients failing and experiencing treatment-limiting toxicity from co-trimoxazole.

Extrapolation of these audit findings to other centres should be done with caution because this audit is from a single centre and is retrospective, and the criteria for determining treatment-limiting toxicity are at the discretion of the treating clinician and may have varied with time.

In summary, co-trimoxazole when used as treatment for first-episode PCP was effective therapy; however, only 64% of patients completed treatment. Regimen change due to toxicity (28%) was not associated with excess mortality. Treatment failure (7%) was associated with more disease-related complications and a worse outcome.

References

Wharton

, Coleman

, Wofsy

, Trimethoprim-sulfamethoxazole or pentamidine for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. A prospective randomized trial. Ann Intern Med 1986;105:37–44

Sattler

, Cowan

, Nielsen

, Ruskin

. Trimethoprim-sulfamethoxazole compared with pentamidine for treatment of Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome: a prospective, noncrossover study. Ann Intern Med 1988;109:280–7

Klein

, Duncanson

, Lenox

, Trimethoprim-sulfamethoxazole versus pentamidine for Pneumocystis carinii pneumonia in AIDS patients: results of a large prospective randomized treatment trial. AIDS 1992;6:301–5

Jaffe

, Abrams

, Ammann

, Lewis

, Golden

. Complications of co-trimoxazole in treatment of AIDS-associated Pneumocystis carinii pneumonia in homosexual men. Lancet 1983;2:1109–11

Hughes

, Lafon

, Scott

, Masur

. Adverse events associated with trimethoprim-sulfamethoxazole and atovaquone during the treatment of AIDS-related Pneumocystis carinii pneumonia. J Infect Dis 1995;171:1295–301

Gordin

, Simon

, Wofsy

, Mills

. Adverse reactions to trimethoprim-sulfamethoxazole in patients with the acquired immunodeficiency syndrome. Ann Intern Med 1984;100:495–9

Benfield

, Atzori

, Miller

, Helweg-Larsen

. Second-line salvage therapy of AIDS-associated Pneumocystis jirovecii pneumonia. J Acquir Immune Defic Syndr 2008;48:63–67

Walzer

, Evans

HER

, Copas

, Edwards

, Grant

, Miller

. Early predictors of mortality from Pneumocystis jirovecii pneumonia in HIV-infected patients: 1985–2006. Clin Infect Dis 2008;46:625–33

Gill

, Maggs

, Madden

, Pirmohamed

, Park

. Co-administration of fluconazole (but not ketoconazole) to healthy volunteers inhibits oxidation of sulphamethoxazole to the hydroxylamine metabolite. Br J Clin Pharmacol 1996;42:347–53

10.

National Institutes of Health (NIH), the Centers for Disease Control and Prevention (CDC), and the HIV Medicine Association of the Infectious Diseases Society of America (HIVMA/IDSA). Guidelines for prevention and treatment of opportunistic infections in HIV-infected adults and adolescents (2008). http://aidsinfo.nih.gov/contentfiles/Adult_OI.pdf (last accessed 18 November 2008)