DOI: https://doi.org/10.21203/rs.3.rs-1601287/v1
Background: Pneumocystis jirovecii pneumonia (PJP) is a life-threatening disease with increasing prevalence in patients with rheumatic disease. Trimethoprim/sulfamethoxazole (TMP/SMX) is an effective treatment for patients with rheumatic disease hospitalized for PJP. This study aimed to describe the 90-day mortality of patients with rheumatic disease complicated by PJP and investigate whether the administration of TMP/SMX after 7 days from initial symptoms correlates with 90-day mortality.
Methods: We enrolled consecutive patients with rheumatic disease complicated with PJP in our center from August 2018 to August 2021. The participants were classified into two groups according to when TMP/SMX was initiated: early (within the first 7 days) and late (after 7 days). The primary outcome was 90-day PJP-related mortality. Multivariate cox regression and Kaplan–Meier survival analyses were conducted to identify the risk factors for mortality and examine differences in survival between early and late use of TMP/SMX.
Results: Thirty-seven patients with rheumatic disease (median age 50.1 years, 24.3% male) complicated by PJP were enrolled in our study, and 15 (40.5%) patients died at or before 90 days of follow-up. The most common comorbidity was systemic lupus erythematosus (14, 37.8%), followed by inflammatory myopathy (11, 27.9%). Patients in the early group were less likely to require mechanical ventilation (29.6% vs. 90.0%, P = 0.002), and had lower mortality (25.9% vs. 80.0%, P = 0.006) than those in the late group. In the Kaplan–Meier analysis, the survivor probability of the early group was notably higher than that of the late group (P = 0.007). Multivariate cox regression analysis showed that initiation of TMP/SMX after 7 days from admission (hazard ratio [HR]: 5.9, 95% confidence interval [CI]: 1.1-30.4; P = 0.034) and a higher level of lactate dehydrogenase (LDH; HR: 6.0, 95% CI: 1.1-31.8; P = 0.035) were associated with 90-day mortality in patients with rheumatic disease complicated by PJP.
Conclusion: Patients with rheumatic disease complicated by PJP had poor prognoses, with mortality rates as high as 40.5%. TMP/SMX initiation after 7 days from initial symptoms and a higher level of serum LDH were significantly associated with increased 90-day mortality.
Pneumocystis jirovecii pneumonia (PJP) is caused by an opportunist fungus that may cause severe or even fatal outcomes in immunocompromised patients, with an incidence approaching 400,000 cases per year globally[1]. Although PJP was initially identified during the acquired immune deficiency syndrome epidemic, its mortality has declined significantly over the past decades due to the use of highly active antiretroviral therapy[2]. In contrast, it has emerged as a significant cause of a typical pneumonia among non-human immunodeficiency virus (HIV) immunocompromised patients who received steroids and cytotoxic drugs[3], the principal therapeutic agents for many rheumatic diseases, which include systemic lupus erythematosus (SLE), inflammatory myopathy (IM), Sjogren's syndrome, Adult-onset Still's disease, Anti-synthetase syndrome, and rheumatoid arthritis.
The precise incidence of PJP in rheumatic diseases is unknown, but a previous study depicted a substantially higher prevalence in patients with SLE, inflammatory myositis, and Wegener’s granulomatosis[4]. More importantly, PJP in patients with rheumatic diseases and non-HIV tends to manifest more severe symptoms and worse prognoses, paralleled by higher oxygen requirements and the need for mechanical ventilation[5] [6] [7]. Effective treatments for PJP in rheumatic diseases are therefore of paramount clinical importance.
Trimethoprim/sulfamethoxazole (TMP/SMX) has been considered the first-line therapy for treating PJP, according to national and international guidelines[8]. However, in clinical practice, the potential harm caused by TMP/SMX administration for patients with rheumatic diseases is also significant, such as drug allergy, renal dysfunction, and thrombocytopenia[9].
Extensive ground-glass opacity is the most typical radiographic finding in PJP[10] on high-resolution computed tomography (CT), which resembles interstitial lung lesions of patients with rheumatic diseases. Therefore, no delay in diagnosing PJP and prescribing TMP/SMX can be challenging due to the nonspecific nature of clinical and radiographic findings, as well as the concerns about adverse effects of the drugs. Recently, it has been shown that prophylactic TMP/SMX dramatically reduces the incidence of PJP in patients with rheumatic diseases who received prolonged, high-dose steroid treatment[11]. However, it is uncertain whether early TMP/SMX administration can improve outcomes in patients with rheumatic disease with definite PJP diagnoses.
We performed this study to describe the risk factors associated with mortality in patients with rheumatic disease complicated by PJP and further assess whether initiating treatment with TMP/SMX within 7 days is independently associated with better clinical outcomes, such as lower rates of mortality and the need for mechanical ventilation, than initiating TMP/SMX after 7 days from initial clinical symptoms.
We retrospectively reviewed the medical records of patients with rheumatic diseases in our center with discharge diagnoses of PJP in Shanghai Jiao Tong University of Medicine affiliated with Renji Hospital from August 2018 to August 2021. Every patient was continuously followed up for 3 months in our center's inpatient and outpatient departments using medical records or occasionally by phone. All participants’ death certificates were reviewed to ascertain the causes of death. It is a retrospective study based on data collected from patients’ clinical records, approved by the Ethics Committee of the Shanghai Jiaotong University School of Medicine Affiliated with Renji hospitals. Informed consent was waived by the Ethics Committee of the Shanghai Jiaotong University School of Medicine Affiliated with Renji hospitals due to the retrospective and observational nature of the study. The data used in this study was anonymised before its use. All methods were performed in accordance with the relevant guidelines and regulations.
The diagnosis of PJP were made according to clinical manifestations, imaging examinations, and microbiological test results. The criteria were as follows: (1) compatible clinical symptoms including fever, cough, sputum, and dyspnea; (2) radiological findings compatible with PJP such as uni- or bilateral ground-glass opacity or patchy consolidation; and (3) etiological basis via molecular diagnoses, such as a positive polymerase chain reaction results for P. jirovecii on samples of sputum or next-generation sequencing (NGS) of serum and bronchoalveolar lavage fluid (BALF). The definitive diagnosis of PJP required identification of the organism. An experienced physician, radiologist, and microbiologist confirmed the diagnoses of all cases.
We excluded patients with any of the following criteria: (1) insufficient data or loss to follow-up; (2) those receiving TMP/SMX prophylaxis.
All data were obtained from the patients’ electronic and physical medical records, and two trained observers collected anonymous patient data using a pre‑specified case report form. Patients' information was collected, including demographics, types and history of rheumatic diseases, types of immunosuppressive agents, time from initial symptom to TMP/SMX use, comorbidities, laboratory results, and clinical outcomes. The primary outcome of the study was 90-day PJP-related mortality.
Continuous variables are presented as means and standard deviations, whereas categorical variables are presented as frequencies and percentages. For the univariate analysis, categorical variables were compared using the Chi‑square or Fisher’s exact tests. The Mann–Whitney U test was used for analyzing continuous variables as they were non-normally distributed. Uni- and multivariate cox regression analyses were conducted to identify the risk factors for 90-day PJP-related mortality. The cut-off of 7 days from initial symptom to TMP/SMX use was selected based on real-world practice. The cut-off of lactate dehydrogenase (LDH) level was selected based on published literature[12, 13]. The variables entered into multivariate analysis were chosen from the factors with P < 0.02 in the univariate analysis. Point estimates are presented as hazard ratios (HRs) with 95% confidence intervals (CIs). The association between TMP/SMX administration within 7 days and mortality was tested by the log-rank test using Kaplan–Meier survival curves. All analyses were performed using SPSS version 23.0.0.0 (SPSS, IBM, USA). P-values < 0.05 were considered statistically significant.
Data of 46 patients with rheumatic disease who met the inclusion criteria were collected between August 2018 and August 2021. Finally, 37 patients were enrolled, 2 were excluded due to severe data loss, 3 were excluded due to loss to follow-up and 4 were excluded because of receiving TMP/SMX prophylaxis.
Baseline characteristics include age, sex, underlying rheumatic diseases, comorbidities, laboratory results, co-infections, and pre- and in-hospital treatment. We divided the patients into the early and late groups based on the time to TMP/SMX initiation whether fewer or > 7 days after administration, respectively. Baseline characteristics and clinical outcomes between the two groups are shown in Table 1. The median age of patients with rheumatic disease with PJP was 50.1 years old, and 75.7% were female. There were 27 (73.0%) patients to whom TMP/SMX was administered within 7 days. The primary comorbidity was SLE (14 patients, 37.8%), followed by IM (11 patients, 27.9%). Cyclophosphamide was used in 24.3% of patients, and 21.6% received more than two immunosuppressive agents. Cytomegalovirus (CMV) (16,43.2%) was the most common co-infection pathogen. Fifteen patients (40.5%) died at 90 days of follow-up. There was no significant difference in baseline data between the two groups. Nevertheless, patients treated with TMP/SMX after 7 days from initial symptoms more likely required mechanical ventilation (90.0% vs. 29.6%, P = 0.002) and had higher 90-day mortality rate (80.0% vs. 25.9%, P = 0.006).
Variables | Total (n = 37) | TMP-SMX ≤ 7days1 (n = 27) | TMP-SMX > 7days2 (n = 10) | P-value |
---|---|---|---|---|
Age (years), mean ± SD | 50.1 ± 13.4 | 52.3 ± 13.2 | 43.9 ± 12.5 | 0.105 |
Sex (male), n (%) | 9 (24.3) | 7 (25.9) | 2 (20.0) | 0.999 |
Underlying rheumatic diseases, n (%) | ||||
Systemic Lupus Erythematosus | 14 (37.8) | 9 (33.3) | 5 (50.0) | 0.454 |
Dermatomyositis | 11 (29.7) | 6 (22.2) | 5 (50.0) | 0.125 |
Others3 | 12 (32.4) | 12 (44.4) | 0 (00.0) | NA4 |
Comorbidity, n (%) | ||||
Hypertension | 7 (18.9) | 4 (14.8) | 3 (30.0) | 0.360 |
Diabetes mellitus | 4 (10.8) | 3 (11.1) | 1 (10.0) | 0.999 |
Heart Failure | 6 (16.2) | 4 (14.8) | 2 (20.0) | 0.653 |
Chronic Renal Dysfunction | 10 (27) | 7 (25.9) | 3 (30.0) | 0.999 |
Interstitial Pneumonia | 17 (45.9) | 10 (37.0) | 7 (70.0) | 0.136 |
Laboratory data, median (range) | ||||
LDH (U/L) | 532 (447–794) | 490 (445–685) | 667(505–1267) | 0.105 |
Lymphocyte count (109/L) | 0.5 (0.3-1.0) | 0.5 (0.3-1.0) | 0.6 (0.2–0.8) | 0.709 |
Albumin (g/L) | 27.0 (23.4–30.4) | 27.0 (23.5–30.0) | 29.0 (23.1–30.2) | 0.662 |
CD4 + lymphocyte count (cells/µL) | 121.8 (52.0-227.0) | 89.9 (59.0-192.0) | 153.5 (60.8-243.5) | 0.783 |
PaO2/FiO2 | 190.0 (120.0-263.5) | 221.0 (145.0-270.0) | 125.0 (100.0-222.5) | 0.089 |
C reactive protein (mg/L) | 58.7 (22.4-110.1) | 75.99 (25.2-113.9) | 50.0 (15.0-70.5) | 0.183 |
Positive EBV DNA, n (%) | 4 (10.8) | 4 (14.8) | 0 (0.0) | NA |
Positive CMV DNA, n (%) | 16 (43.2) | 13 (48.2) | 3 (30.0) | 0.461 |
Previous treatment5, n (%) | ||||
CTX use | 9 (24.3) | 9 (33.3) | 0 (0.0) | NA |
≧2 immunosuppressive agents use | 8 (21.6) | 7 (25.9) | 1 (10.0) | 0.404 |
Dose of steroids6 (mg/d) , median (range) | 40.0 (28.8–54.0) | 40.0 (30.0–50.0) | 40.0 (21.5–52.5) | 0.857 |
In-hospital treatment, n (%) | ||||
Caspofungin use | 30 (81.1) | 20 (74.1) | 10 (100.0) | 0.155 |
Mechanical ventilation | 17 (45.9) | 8 (29.6) | 9 (90.0) | 0.002 |
90-day PJP-related mortality, n (%) | 15 (40.5) | 7 (25.9) | 8 (80.0) | 0.006 |
1TMP-SMX ≤ 7days: time from initial symptom to TMP/SMX use no more than seven days | ||||
2TMP-SMX > 7 days: time from initial symptom to TMP/SMX use more than seven days | ||||
3Others: including Sjogren's syndrome (pSS), Adult-onset Still's disease (AoSD), Anti-synthetase syndrome (ASS), Rheumatoid Arthritis (RA), Arteritis | ||||
4NA: not applicated | ||||
5Previous treatment: treatment measures before hospitalization | ||||
6Dose of steroids: daily exposure dose of glucocorticoid within three months expressed as the methylprednisolone equivalent dose |
A comparison of clinical baseline data in deceased and surviving patients was presented in Table 2. There were no differences in age, sex, basic rheumatic diseases, comorbidity diseases, co-infection, previous immunosuppressive agents use, the daily exposure dose of glucocorticoid within three months between death and survival group. However, we found higher PaO2/FiO2 ratios (246.5 vs. 120.0, P༜0.001) and serum albumin levels (28.2 g/L vs. 23.8 g/L, P = 0.034), lower levels of LDH (474.5 U/L vs. 685.0 U/L, P = 0.005) and early treatment with TMP/SMX (9.1% vs. 53.3%, P = 0.003) in survivors compared to non-survivors.
Variables | Survival (n = 22) | Death (n = 15) | P-value |
---|---|---|---|
Age (years), mean ± SD | 50.0 ± 13.3 | 50.1 ± 14.1 | 0.977 |
Sex (male), n (%) | 5 (22.7) | 4 (26.7) | 0.784 |
Underlying rheumatic diseases, n (%) | |||
Systemic Lupus Erythematosus | 8 (36.4) | 6 (40.0) | 0.999 |
Dermatomyositis | 6 (27.3) | 5 (33.3) | 0.728 |
Others | 8 (36.4) | 4 (26.7) | 0.724 |
Comorbidity, n (%) | |||
Hypertension | 4 (18.2) | 3 (20.0) | 0.890 |
Diabetes mellitus | 2 (9.1) | 2 (13.3) | 0.683 |
Heart Failure | 3 (13.6) | 3 (20.0) | 0.606 |
Chronic Renal Dysfunction | 5 (22.7) | 5 (33.3) | 0.476 |
Interstitial Pneumonia | 8 (36.4) | 9 (60.0) | 0.157 |
Laboratory, median (range) | |||
LDH (U/L) | 474.5 (403.0-667.0) | 685.0 (535.5–1305.0) | 0.005 |
Lymphocyte count (109/L) | 0.6 (0.30-1.0) | 0.4 (0.2–0.8) | 0.345 |
CD4 + lymphocyte count (cells/µL) | 137.9 (65.6-207.8) | 82.0 (41.5–213.0) | 0.404 |
Albumin (g/L) | 28.2 (26.0-31.5) | 23.8 (21.4–29.6) | 0.034 |
PaO2/FiO2 | 246.5 (192.5–280.0) | 120.0 (105.0-147.5) | < 0.001 |
C reactive protein (mg/L) | 49.9 (23.7–92.8) | 90.7 (36.1-117.9) | 0.337 |
Positive EBV DNA n (%) | 3 (13.6) | 1 (6.7) | 0.503 |
Positive CMV DNA, n (%) | 8 (36.4) | 8 (53.3) | 0.306 |
Previous treatment, n (%) | |||
TMP-SMX > 7 days | 2 (9.1) | 8 (53.3) | 0.006 |
CTX use | 9 (40.9) | 0 (0.0) | NA |
≧2 immunosuppressive agents use | 5 (22.7) | 3 (20.0) | 0.843 |
Dose of steroids (mg/d) , median (range) | 40.0 (30.0–50.0) | 40.0 (24.8–60.0) | 0.596 |
In-hospital treatment, n (%) | |||
Caspofungin use | 16 (72.7) | 14 (93.3) | 0.116 |
Mechanical ventilation | 3 (13.6) | 14 (93.3) | < 0.001 |
Univariate and multivariate analyses of risk factors for mortality are illustrated in Table 3. Multivariate cox analyses showed that patients who were prescribed TMP/SMX after 7 days (HR: 5.9; 95% CI: 1.14–30.43; P = 0.034) and a higher level of LDH (HR: 6.0; 95% CI: 1.13–31.78; P = 0.035) were associated with higher mortality in patients with rheumatic disease complicated by PJP.
Variable | Univariable | Multivariable | ||
---|---|---|---|---|
HR (95% CI) | P-value | HR (95%CI) | P-value | |
TMP-SMX > 7 days | 4.0 (1.4, 11.3) | 0.009 | 5.9 (1.1, 30.4) | 0.034 |
Age, per 1 years old | 1.0 (1.0, 1.0) | 0.827 | ||
Sex (female or male) | 1.2 (0.4, 3.8) | 0.757 | ||
Interstitial Pneumonia (yes or no) | 1.9 (0.7, 5.5) | 0.210 | ||
LDH > 500 U/L | 7.2 (1.6, 32.1) | 0.010 | 6.0 (1.1, 31.8) | 0.035 |
Lymphopenia < 1000 cells/µL | 0.8 (0.2, 2.7) | 0.660 | ||
CD4 + lymphocyte count < 200 cells/µL | 1.0 (0.3, 3.0) | 0.947 | ||
Albumin < 30g/L | 1.7 (0.5, 6.2) | 0.393 | ||
Caspofungin (yes or no) | 3.7 (0.5, 28.0) | 0.210 | ||
Dose of steroids, per 10 mg/d | 1.9 (0.5, 6.7) | 0.353 |
Kaplan–Meier survival analysis in Fig. 1 shows that patients who received early treatment with TMP/SMX within 7 days had a better prognosis than those who did not receive TMP/SMX within 7 days (P = 0.007).
This study reports that patients with rheumatic disease complicated by PJP who received TMP/SMX > 7 days after initial symptoms had an increased risk of 90-day mortality compared with those who received TMP/SMX within 7 days. Additionally, a higher level of LDH also correlated with a worse prognosis in patients with rheumatic diseases complicated by PJP.
Previous studies have shown that long-term exposure to glucocorticoid and immunosuppressant therapy increases the risk of PJP infection in patients with rheumatic diseases[14]. Due to the immunosuppressive condition of patients, they are prone to have co-infection, which may be exacerbated death[14] [15] [16]. Similar to previous studies[12], in our research, the 90-day mortality rate of PJP patients was as high as 40.5%, and the evidence of mixed infection such as aspergillosis, CMV, and Epstein–Barr virus was also found in the NGS of blood and BALF samples.
The characteristic presentation of PJP included acute fever, dyspnea, nonproductive cough, hypoxemia, and even thoracic pain in clinical settings[13]. PJP presented as bilateral interstitial infiltrates and bilateral ground-glass attenuation on thoracic CT scans[17]. PJP should be highly suspected when immunocompromised patients with these typical clinical symptoms and CT signs[18], and prompt antimicrobial treatment against P. jirovecii should be triggered[5] [19] [20]. With the promotion of NGS application in clinical settings, the detection rate of P. jirovecii is significantly increasing[21] [22]. Our study primarily made definite PJP diagnoses by positive NGS assays from BALF specimens and occasionally by hypertonic saline-induced sputum smears.
Since delayed treatment could increase mortality and the need for mechanical ventilation, immediate initiation of PJP-specific therapy is critical[5, 19, 20]. Based on the real-world clinical setting, for patients suspected of PJP, 1 week was considered the acceptable cut-off point for TMP/SMX preemptive treatment, so we divided the patients into two groups to be compared. The baseline characteristics of patients in the two groups were comparable; however, the outcomes differed even after multivariate analyses. Our research supports the efficacy of preemptive TMP/SMX administration as evidenced by reduced mortality and improved prognoses. Cox regression analysis showed that initiation of TMP/SMX after > 7 days from the onset of clinical symptoms was related to higher 90-day all-cause mortality than those who received TMP/SMX treatment within 7 days. Our study reconfirms that prompt treatment with TMP/SMX is a critical measure to reduce mortality among patients with rheumatic disease complicated with PJP.
Previous clinical studies have shown that high LDH levels were a risk factor for adverse outcomes in PJP patients[23]. A clinical retrospective study over 17 years found that initial LDH levels can be used as a stratification tool to identify patients with a high risk of mortality[24]. In addition, the prolonged usage of glucocorticoids, delayed anti-P. jirovecii therapy, hypoalbuminemia, CMV infection, acute respiratory distress syndrome, severe pulmonary edema, respiratory failure, the need for mechanical ventilation, and high-dose glucocorticoid therapy during hospitalization were all predictive of poor outcomes[18] [25] [26] [27] [28] [29] [30].
This study demonstrated that a higher level of LDH and lower PaO2/FiO2 ratio in the early stage of PJP resulted in adverse outcomes. Consistent with previous studies, the multivariate analyses showed that higher level LDH was associated with a worse prognosis in PJP patients with rheumatic diseases.
The study has several limitations. First, data loss is unavoidable due to the retrospective experimental design, such as the loss of BALF cell counts, which leads to the loss of some indicators of PJP prognoses. Second, the study's sample size was limited due to the disease’s rarity. Thirdly, our single-center retrospective design reduced the generalizability of our findings. In the future, we will undertake high-quality multi-center studies to analyze the risk factors affecting the prognosis of PJP patients with rheumatic diseases.
Patients with rheumatic disease complicated by PJP are associated with poor prognoses regardless of specific anti-P. jirovecii treatment. Time to TMP/SMX initiation > 7 days after initial symptoms and higher serum LDH levels were associated with adverse outcomes in patients with rheumatic disease with PJP.
BALF, bronchoalveolar lavage fluid
CI, confidence interval
CT, computed tomography
CMV, cytomegalovirus
HIV, human immunodeficiency virus
HR, hazard ratio
IM, inflammatory myopathy
LDH, lactate dehydrogenase
NGS, next-generation sequencing
PJP, Pneumocystis jiroveci pneumonia
SLE, systemic lupus erythematosus
TMP/SMX, trimethoprim/sulfamethoxazole
The findings and conclusions in this study are those of the authors and do not necessarily represent the official position of the funding agencies.
Ethics approval and consent to participate
It is a retrospective study based on data collected from patients’ clinical records, approved by the Ethics Committee of the Shanghai Jiaotong University School of Medicine Affiliated with Renji hospitals. Informed consent was waived by the Ethics Committee of the Shanghai Jiaotong University School of Medicine Affiliated with Renji hospitals due to the retrospective and observational nature of the study. The data used in this study was anonymised before its use. All methods were performed in accordance with the relevant guidelines and regulations.
Consent for publication
Not applicable.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing interests
All authors declare no conflict of interest.
Funding
Not applicable.
Authors’ contributions
YZ and JY collected all the clinical data; SS and XZ were involved in statistical data analysis and editing the manuscript; YC and CX designed the study. All authors read and approved the final manuscript.
Acknowledgements
We wish to thank our colleagues at Renji Hospital for their help and advice, Dr. Wanglong Wu for his expert advice assistance and Dr. Chenyu Fan for his technical assistance in this study.
Authors' information
Department of Emergency, Shanghai Jiaotong University School of Medicine affiliated Renji Hospital, 2000 Jiangyue Road, Minhang District, Shanghai, China