FormalPara Key Points

Why carry out this study?

Pneumocystis pneumonia is a common opportunistic infection in patients with HIV/AIDS, and is a leading cause of death in this population. Early selection of effective treatment is therefore critical to reduce mortality.

Some patients show poor sensitivity to TMP-SMX monotherapy. TMP-SMX may be slow to achieve curative effect, which suggests that this treatment option may be inappropriate for critically ill patients.

Primaquine plus clindamycin has been suggested in the current US Department of Health and Human Services (DHHS) guideline as an alternative treatment for patients who have moderate to severe Pneumocystis pneumonia (PCP). Because primaquine is not available in China owing to the elimination of malaria there, it has become necessary to seek a new therapeutic regimen to replace primaquine.

Caspofungin plus TMP-SMX may be a promising drug combination for use in patients with HIV/PCP. However, clinical experience with use of this combination is currently limited.

The aim of this study was to compare the effectiveness and safety of three different antifungal treatment regimens in HIV-infected patients with moderate to severe PCP.

What was learned from the study?

TMP-SMX monotherapy as a therapeutic drug regimen to treat HIV-infected patients with moderate to severe PCP is an appropriate treatment strategy in resource-limited settings.

Introduction

Pneumocystis pneumonia (PCP) is caused by Pneumocystis jirovecii, which was previously classified as a protozoan, but is currently considered to be a fungus on the basis of nucleic acid and biochemical analysis. PCP is one of the commonest opportunistic infections in patients with HIV/AIDS, often occurring in patients with CD4+ T cell counts less than 200 cells/μL, and is one of the main causes of hospitalization and death among patients with HIV/AIDS. Some early studies showed that the mortality of HIV-infected patients with severe PCP was 36–50% in the early antiretroviral therapy (ART) era [1]. A past study showed that in patients with HIV/PCP who had received ART, mortality was 9.9%; in those who had not received ART, mortality was 12.0% [2]. In the modern ART era, the mortality rate in HIV-infected patients with PCP has decreased to approximately 10–12%; however, mortality can rise to up to 84% in those with moderate to severe disease without any treatment [3]. Therefore, early selection of effective treatment is a priority in these patients.

Trimethoprim/sulfamethoxazole (TMP-SMX), also known as co-trimoxazole/Bactrim®, is currently recommended as first-line treatment of PCP in HIV-infected patients because of its relatively high overall efficacy and the widespread availability of its oral and parenteral formulations [1]. TMP-SMX can relieve the symptoms of PCP and improve prognosis. However, the efficacy rate of patients on TMP-SMX monotherapy administered to HIV-infected patients with moderate to severe PCP is 70–80%, and some patients show poor sensitivity to TMP-SMX monotherapy [4]. TMP-SMX is considered to only destroy trophoblasts, and has no effect on the tomont stage of the organism [5]. In addition, TMP-SMX is slow to act, requiring approximately 5–8 days to achieve curative effect, which suggests that the use of TMP/SMX may be inappropriate for critically ill patients [6]. Therefore, the quest for more effective novel therapeutic regimens to improve the prognosis of patients with moderate to severe PCP is an important and currently relevant issue to be addressed [7].

Clindamycin belongs to the lincosamine class of antibiotics, and can act as an alternative to TMP-SMX in terms of therapeutic success and patient safety [8]. Used together with primaquine, clindamycin has been suggested for patients who have moderate to severe PCP in the current US Department of Health and Human Services (DHHS) guideline [9]. Because primaquine is not available in China owing to the elimination of malaria there, it has become necessary to seek a new therapeutic option to replace it. Not many reports in the literature exist with regards to the combination of TMP-SMX with clindamycin [10], even though in clinical practice the efficacy of TMP-SMX plus clindamycin appears to be greater than that of TMP-SMX monotherapy.

In recent years, an increasing number of studies report that caspofungin may be used to treat PCP [11]. However, it is believed at present that caspofungin alone should not be used for the treatment of PCP (especially in moderate to severe PCP), as it is unlikely to completely clear Pneumocystis, and may allow PCP recurrence after drug withdrawal [12]. TMP-SMX exerts its effects on trophic forms of P. jirovecii, while caspofungin primarily acts on cystic forms of P. jirovecii [13], and the combination of caspofungin and TMP-SMX thus has the potential to inhibit the entire life cycle of the Pneumocystis organism [14]. Case reports have described the clinical success of caspofungin when used in combination with TMP-SMX in treating moderate to severe PCP in immunocompromised hosts [15]. A clinical trial involving caspofungin salvage treatment in HIV-infected patients with PCP has shown an appreciably high success rate (80%, 8/10 patients) [13]. Jin et al. found that HIV-negative patients with moderate to severe P. jirovecii pneumonia had good response to caspofungin combined with TMP-SMX [16]. In further studies, four patients with PCP were successfully treated with caspofungin and TMP-SMX [17]. Caspofungin may thus be a promising drug for use in patients with HIV/PCP; however, clinical experience with its use is currently limited.

In our multicenter, prospective observational cohort study of data from 16 hospitals throughout China, we aimed to observe and compare the clinical effectiveness and safety of TMP-SMX monotherapy, TMP-SMX plus clindamycin, and TMP-SMX plus caspofungin for the treatment of moderate to severe PCP (defined by the presence of an arterial oxygen pressure of less than 70 mmHg or an arterial-alveolar gradient of more than 35 mmHg in a patient with HIV/PCP) in HIV-infected patients.

Methods

Patient Population and Study Design

The present multicenter, observational cohort study was conducted at 16 hospitals in China, from January 2019 to December 2020. Our study received human research ethics approval (Approval No. 2019-003-02-KY) from the Ethics Committee of the Chongqing Public Health Medical Center, and from the individual institutional ethics committees of each of the other 15 hospitals involved in this study (Table S1 in the supplementary material), and was duly registered at the Chinese Clinical Trial Registry (Registration No. ChiCTR1900021195). The study was conducted in accordance with the tenets of the 1964 Declaration of Helsinki. All enrolled patients provided written informed consent.

Inclusion and Exclusion Criteria

Patients with confirmed PCP were included in our study if they satisfied the following eligibility criteria: (1) over 18 years old; (2) HIV positive; (3) willing to sign informed consent; (4) moderate-to-severe PCP.

Patients were excluded from this study if they (1) were intolerant of or had severe allergy to any of the therapeutic drugs used; (2) had hemoglobin (HGB) levels less than 60 g/L, white blood cell counts (WBC) less than 1.0 × 109/L, neutrophil counts (N) less than 0.5 × 109/L, platelet counts (PLT) less than 50 × 109/L, blood amylase (AMS) more than two times the upper normal limit (UNL), serum creatinine (Scr) more than 1.5 times UNL, aspartate aminotransferase (AST)/alanine aminotransferase (ALT)/alkaline phosphatase (ALP) more than five times UNL, total bilirubin (TBIL) more than two times UNL, serum creatine phosphokinase (CK) more than two times UNL; (3) had the presence of other serious disease that could have affected the accurate evaluation of efficacy and prognosis; (4) were pregnant or breast-feeding women; (5) had severe mental health illnesses, or used intravenous recreational drugs, (6) had non-Chinese nationality; (7) had mild PCP; or (7) withheld informed consent.

Definitions

Moderate to severe of PCP was defined as: (1) partial arterial oxygen pressure (PaO2) ≤ 70 mmHg, or while breathing room air, an alveolar-arterial oxygen difference of (A-aDO2) ≥ 35 mmHg; (2) the presence of relevant pulmonary symptoms, including dry cough, shortness of breath, progressive dyspnea, and purpura, and may also have fever; (3) pulmonary infiltrations indicated by chest radiography or computed tomography (CT) performed when PCP was clinically suspected following hospitalization; (4) microbiological confirmation by positive PCR and/or Grocott’s methenamine silver (GMS) results for sputum, pulmonary aspirate, or broncho-alveolar lavage fluid (BALF) samples. If the subject met criteria (1), and (2), and (3), and/or (4), moderate to severe of PCP was assumed to be present. PaO2 was measured directly via blood gas analysis of arterial blood.

Overall positive response rate was defined as having fewer clinical symptoms, improved PaO2, and resolution of evidence of pneumonitis on chest imaging after treatment.

Treatment failure was defined as persistent fever and worsening hypoxia, and/or radiographic deterioration.

Data Collection and Quality Assurance

Based on a standard operating procedure (SOP) manual, investigators were trained to ensure patient adherence to the study protocol. Data including clinical characteristics, efficacy, safety, and adverse effects were collected at each follow-up visit during the follow-up period. All raw data were recorded in case report forms (CRFs) and then transferred to an electronic database via the Medical Research Platform. Withdrawals from the study, or missed visits were fully explained on CRFs. The study monitor reviewed the completed CRFs quarterly to ensure accuracy and diligence of the application of inclusion, exclusion, and withdrawal criteria, as well as to ensure that information on the CRFs were consistent with those in the source electronic medical records.

Treatment and Outcomes

All eligible patients were divided into three groups: Group 1, receiving TMP-SMX monotherapy, Group 2, receiving TMP-SMX plus clindamycin, and Group 3, receiving TMP-SMX plus caspofungin. During the study period, we used similar protocols for drug dosage and administration as recommended by international guidelines. TMP-SMX was given at a daily dose of 15–20 mg/kg of trimethoprim and 75-100 mg/kg of sulfamethoxazole. Clindamycin was given by intravenous injection at dose of 0.6 g every day. Patients received 70 mg caspofungin intravenously on the first day, and were then subsequently administered a dose of 50 mg/day. All treatment regimens lasted for 3–4 weeks. Adjunctive corticosteroids at the dose recommended by guidelines were administered to all patients. Patients received additional nasal cannula oxygen therapy, and non-steroidal anti-inflammatory drugs (NSAIDs) as required, as well as oral or intravenous rehydration, electrolyte correction, antipyretics, analgesics, and antiemetic drugs as their individual clinical conditions demanded.

The primary outcome was the difference in overall mortality in the three groups at week 4 and week 12. Secondary outcomes included the proportion of overall positive response to treatment of moderate to severe PCP in each group at week 4 and week 12, the differences among the three groups in the duration of treatment, and the difference in rates of adverse events among the three groups during the study period.

Study Procedures

After HIV/PCP patients were admitted to hospital, we selected those patients whom we assessed may be eligible for our study, and sought their consent for eligibility screening. After they provided their informed consent, we assessed their eligibility for study inclusion, and only included into our study those who met our inclusion criteria but who also did not meet the exclusion criteria of our study. Study visits occurred at week 1, week 2, week 3, week 4, and week 12 after initiation of PCP treatment. Safety was assessed by interrogation of the participants for potential development of clinical symptoms, and physical examination for clinical signs, clinical laboratory tests, and documentation of adverse events. Treatment adherence was assessed by review of clinical diaries that were filled out by medical staff.

Statistics

Continuous variables were described as mean (± SD) for normally distributed data, or median with interquartile ranges (IQR) for non-normally distributed data. Categorical variables were expressed as frequency rates and percentages. Kolmogorov–Smirnov tests and ANOVA (analysis of variance) tests were used for continuous variables. Chi-squared tests, continuity corrections, and Fisher’s exact tests were used to test statistical significance for categorical data. Cox proportional-hazards model were used to analyze and compare survival rates. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) software, version 25.0 software (IBM SPSS, Armonk, New York, USA), using a statistical significance threshold of p < 0.05.

Sample size was calculated by using the PASS (Power Analysis and Sample Size) software, version 15 (NCSS, LLC, Kaysville, Utah, USA), according to the following assumptions: a 12-week mortality rate of 30%, 15%, and 10% in the TMP-SMX monotherapy group, the TMP-SMX plus clindamycin group, and the TMP-SMX plus caspofungin group, respectively, with an overall two-sided alpha level of 0.05, and a statistical power of 90%. Thus, a sample size of 263 patients was required for our study. Considering a 10% withdrawal rate, we planned to enroll 300 participants, and thus have at least 270 participants for analysis.

Results

Baseline Characteristics

During the study period, 356 HIV-infected patients with moderate to severe PCP were screened for eligibility, of whom 320 patients were assigned to an intervention group: 123 patients were assigned to group 1 (the TMP-SMX monotherapy group), 87 patients were assigned to group 2 (the TMP-SMX plus clindamycin group), 110 patients were assigned to group 3 (the TMP-SMX plus caspofungin group) (Fig. 1).

Fig. 1
figure 1

Flow diagram of the study

Full size image

The mean age of the cohort was 47.46 years old, and 250 (78.13%) of the 320 patients were men. Most did not have previous exposure to antiretroviral therapy (95.63%, 306/320), 35.63% (114 /320) of the patients had a history of smoking, and 20.94% (67/320) had a history of alcohol use. The most common comorbid diseases present in our cohort were tuberculosis (10.63%, 34/320), diabetes mellitus (3.13%, 10/320), and hypertension (1.56%, 5/320). Because of the presence of hypoxia and dyspnea in moderate to severe PCP, it was technically difficult and potentially hazardous to carry out bronchoscopic examination in most patients. Approximately 10% (32/320) of our patients underwent bronchoscopic examination and bronchoalveolar lavage for definitive PCP diagnosis, comprising 14 patients in group 1, 10 patients in group 2, and 8 patients in group 3.

Table 1 shows the clinical characteristics of patients in groups 1–3. There were no significant differences in age, sex, BMI, HIV RNA levels, G test, ALT levels, AST levels, PaO2 levels, HGB levels, and LDH levels among the three groups. Compared with patients in groups 2 and 3, more patients in group 1 were smokers. The clinical symptoms of PCP on admission included fever, dyspnea, and cough. All patients underwent chest CT scans, and the most common radiological finding was bilateral diffuse ground glass opacification or bilateral predominant consolidation.

Table 1 Baseline characteristics of the enrolled patients
Full size table

All-Cause Mortality

Adjusted for route of infection and smoking, no significant differences in the 12-week time distribution of survival were found in group 2 (p = 0.093, HR 1.760, 95% CI 0.911, 3.403) and group 3 (p = 0.074, HR 1.791, 95% CI 0.945, 3.392), compared with group 1 (Fig. 2). Among all 320 patients investigated in the study period, group 1 exhibited an all-cause mortality rate of 13.04% (15/115), group 2 exhibited an all-cause mortality rate of 24.10% (20/83), while group 3 exhibited an all-cause mortality rate of 22.43% (24/107). However, these apparent differences in all-cause mortality at week 12 were not statistically significant (p = 0.092; Table 2).

Fig. 2
figure 2

Overall survival among all 320 participants for three groups. Group 1, TMP-SMX monotherapy group; group 2, TMP-SMX plus clindamycin group; group 3, TMP-SMX plus caspofungin group

Full size image
Table 2 All-cause mortality rates of the study participants
Full size table

Positive Response Rates

The duration of anti-PCP treatment was 21 days in each of the three groups. No statistically significant difference in treatment duration was observed among the three groups, and between any two of the three groups (p = 0.986) (Fig. 3).

Fig. 3
figure 3

Duration of moderate to severe PCP treatment. Group 1, TMP-SMX monotherapy group; group 2, TMP-SMX plus clindamycin group; group 3, TMP-SMX plus caspofungin group

Full size image

The overall positive response rate to PCP treatment was 24.14% (28/116), 34.94% (29/83), and 38.32% (41/107) in group 1, group 2, and group 3, respectively, at week 4 (p = 0.061). The overall positive response rate to PCP treatment did not differ significantly among the three groups. The positive response rates of the three groups at week 12 were 33.91% (39/115), 38.55% (32/83), and 44.86% (48/107), respectively (p = 0.246). However, no statistically significant difference in the overall positive response rate to PCP treatment was observed among the three groups (Table 3).

Table 3 Positive response rates in each treatment group
Full size table

Safety Evaluation

Changes to drug regimens were classified as either caused by adverse effects or caused by suspected failure of treatment. Of 14 of 320 (4.38%) patients where a specific therapeutic drug was changed to another drug, 5 (35.71%) regimens were changed because of adverse effects and 9 (64.29%) were changed secondary to treatment failure. Drug changes occurred more frequently in group 1 (6.50%), compared with group 2 (3.40%) and group 3 (2.70%); however, the rates of drug changes were not significantly different among the three groups (Table 4). The median time to change was 12 days in group 1, 24 days in group 2, and 7 days in group 3, regardless of the reason for change.

Table 4 Treatment for moderate to severe PCP and reasons for switching treatment
Full size table

Among the three treatment groups, adverse effects mainly presented as bone marrow suppression (p = 0.998, p = 0.988) and electrolyte disturbances (p = 0.151, p = 0.312), and there was no significant difference among the three groups in the occurrence of any adverse event at both week 4 or 12. Three patients had gastrointestinal symptoms (two patients in group 1, one patient in group 2, and none in the group 3), with no statistical difference in the occurrence of gastrointestinal symptoms among the three groups at both week 4 (p = 0.494) or week 12 (p = 0.494). In addition, only one patient appeared to develop central nervous system symptoms (in a patient in the TMP-SMX monotherapy group) in our study. Furthermore, we found no statistically significant differences in the rates of skin rash (p = 0.611) and renal dysfunction (p = 0.590) among three groups at week 12 (Table 5).

Table 5 Cumulative adverse events of regimens at weeks 4 and 12
Full size table

Discussion

Currently, the first choice of therapeutic drug for the prevention and treatment of PCP is TMP-SMX. Studies have shown that the survival rates of HIV-infected patients with PCP are only about 70% when using TMP-SMZ alone as first-line therapy, and survival rates using TMP-SMZ are even lower in moderate to severe cases. Thomas et al. reported an overall survival rate of 93%, and a survival rate of 81% in patients with severe disease, among 73 HIV-infected patients treated with TMP-SMX [18]. Some moderate to severe cases can also prove fatal despite full supportive intensive care with mechanical ventilation. This observation is also supported by the results of our prospective study. In our study, the mortality rate of patients with moderate to severe PCP and HIV was 13.04% in the TMP-SMX monotherapy group. We observed that the TMP-SMX plus clindamycin group and the TMP-SMX plus caspofungin group showed no further decrease in mortality compared with those prescribed TMP-SMX only.

Previous case reports showed that the efficacy of caspofungin was favorable when it was used as a first-line drug or for salvage therapy for PCP [19]. A trial involving the caspofungin salvage therapy for PCP showed a high success rate (80%) among HIV-infected patients [20]. In contrast, Kamboj et al. described four HIV-negative patients with PCP that had no observable clinical response to caspofungin as salvage therapy [21]; however, their study is consistent with ours. We found that both therapeutic combinations (TMP-SMX plus clindamycin and TMP-SMX plus caspofungin) demonstrate similar efficacy for the treatment of moderate to severe PCP as TMP-SMX monotherapy. In recent years, the abuse of antibiotics has increased, which not only facilitates the emergence of antibiotic drug resistance but also results in large amounts of wasted antibiotics and of antibiotics entering the water table and waterways, causes an unnecessary economic burden on patients, and causes unnecessary adverse reactions. Currently, China has come to inherit the unfortunate distinction of becoming one of the world’s worst abusers of antibiotics [22]. Hence, if TMP-SMX monotherapy can be demonstrated to be as effective and efficient as combinations with other antibiotics and antifungals (as has been demonstrated with the results of our study) for the treatment of moderate to severe PCP, then obviously combinations of TMP/SMX with other antibiotics and antifungals should not be used.

The adverse events of TMP-SMX have been widely reported, and include rashes, fever, gastrointestinal complications, bone marrow suppression, hyperkalemia, hepatoxicity, and renal insufficiency [23]. The adverse effects of clindamycin and caspofungin are similar to those of TMP-SMX. In our study, we there was a high incidence of bone marrow suppression, hepatic dysfunction, and electrolyte disturbances, both at 4 weeks and at 12 weeks, in all three groups, and there were no significant differences in adverse effect rates among the three groups, suggesting that combination therapy with TMP-SMX and clindamycin or caspofungin did not increase the incidence of adverse events in subjects compared with TMP-SMX monotherapy. In addition, we failed to identify individual cases of interstitial nephritis, pancreatitis, and aseptic meningitis in our cohort.

Our study had a few limitations. Firstly, not all patients with moderate to severe PCP underwent bronchoscopic examination and bronchoalveolar lavage for definitive PCP diagnosis because hypoxia and dyspnea in some patients made it technically difficult and potentially hazardous to undertake those procedures; nevertheless, we adopted a unified clinical diagnostic standard for patients with moderate to severe PCP who could not undergo those procedures. Secondly, a limited sample size and the non-randomized nature of our study may have introduced a degree of bias into our study. Also, selection and unmeasured confounding bias cannot be completely excluded. Novel therapeutic interventions should ideally be assessed via randomized, controlled clinical trials. However, such an approach may not be practically feasible in the context of patients with critical illness, especially in immunocompromised patients. We endeavored to include patients at a 1:1:1 ratio, so as to reduce inherent bias caused by the subjective intentionality of clinicians in the study. Thirdly, as a result of TMP-SMX drug resistance being rarely reported in China, we did not investigate the occurrence of drug resistance in P. jirovecii in this study.

Conclusion

Our results indicate that there are no significant differences among the three different treatment regimens in terms of antifungal effectiveness in HIV-infected patients with moderate to severe PCP. Caspofungin and clindamycin are expensive, may not be readily available to patients, and are generally used to manage quite specific infectious agents and diseases, while TMP/SMX is inexpensive, readily available, and is commonly used. TMP-SMX monotherapy as a therapeutic drug regimen to treat HIV-infected patients with moderate to severe PCP is therefore an economically viable, convenient, and appropriate treatment strategy in resource-limited settings.