Among the two recruitment sites, 429 participants were selected from Beijing You An Hospital, the famous infectious disease hospital in the capital of the country, and 44 participants were selected from the Infectious Disease Hospital in Harbin, located in northeast China. Although these participants were only selected from the two sites, they were distributed throughout the country, especially North China (Fig. 1). Average age of these subjects was 38.9-year-old, and 95.35 percent of them were men owe to the high HIV infection rate in men. 27.27 percent of subjects had a history of smoking. Over one-third (32.77%) of these patients knew that they had infected HIV before, and 41.29 percent of them had received irregular ART which did not effectively reconstitute host immunity. Most participants were homosexual (30.44%), even if 49.05 percent of them intentionally or unintentionally concealing HIV infection route. Near half of 473 participants accompanied extra-lung organism infection (48.41%), and over two third of them had other organism infection in lung (77.59%). The demographics of two groups were summarized in Table 1. There was no difference in demographics between the two groups except for age and former lung diseases.
2. The clinical feature of these patients with pneumocystis pneumonia
The consciousness of almost all 473 patients was clear even in severe cases (99.63% in mild group and 99.51% in moderate/severe group). The respiratory rates were fast in these patients (21.59±4.75 breath per minute in mild group and 23.18±5.82 breath per minute in moderate/severe group) and the respiratory rates of moderate/severe group were faster than that of mild group (p<0.01). Similarly, heart rates were relatively fast (93.07±15.89 beat per minute in mild group and 95.68±16.10 beat per minute in moderate/severe group). Overall, average auxiliary temperature was 37.14°C in mild group and 37.40°C in moderate/severe group and the body temperature of moderate/severe group were higher than that of mild group (p<0.01). Mean blood pressure of these patients was in the normal range (87.38 mmHg in mild group and 86.91 mmHg in moderate/severe group). The median of white blood cell counts was 5.26×109/L in mild group and 6.67×109/L in moderate/severe group, and white blood cell counts of moderate/severe group was more than that of mild group (p<0.01). These patients had mild anemias and the average hemoglobin were 118.15 g/L in mild group and 119.97 g/L in moderate/severe group. Interestingly, the average platelet count of moderate/severe group (255.95×109/L) was more than that of mild group(220.48×109/L, p<0.01) and the serum creatinine level of moderate/severe group (62.44μmol/L) was lower than that of mild group (67.72μmol/L, p<0.01), which seem to differ from other pneumonia and infectious diseases. The serum lactate dehydrogenase (LDH) and high-sensitivity C-reactive protein (hCRP) of moderate/severe group (LDH 456.82 U/L and hCRP 57.40 mg/L) was higher than that of mild group (LDH 328.65 U/L and hCRP 39.29 mg/L, p<0.01). However, two important parameters of HIV infection, CD4 cell counts and plasma HIV loads were not different between two groups in which CD4 cell counts were 53.53×106/L in mild group vs. 34.19×106/L in moderate/severe group and plasma HIV loads were 5.3 log10 copies/mL in mild group vs. 5.25 log10 copies/mL in moderate/severe group(p>0.05). The mortality of moderate/severe PCP (28.29%) was significantly higher than that of mild PCP (13.06%, p<0.01) (Table2).
3. The strategy and outcome of anti-pneumocystis pneumonia therapy
In this study, 21 days of TMP-SMZ (TMP 15-20 mg/kg/d and SMZ 75-100 mg/kg/d) were selected as standard first-line anti-PCP therapy. Some patients could not tolerate the allergies or various side effects of sulfonamide and had to reduce TMP-SMZ dose or shorten therapeutic course. Here, we did not find the difference in the selection of TMP-SMZ dose or therapeutic course between mild PCP and moderate/severe PCP (Table3). Caspofungin was selected as a combined treatment for the severe cases, or alternative medicine for the cases those could not tolerate side-effect of TMP-SMZ due to lack of second-line anti-PCP medicines including primaquine, pentamidine, dapsone and atovaquone in China. It was shown that combined caspofungin treatment was more common in moderate/severe group (72.20%) than that in mild group (52.61%, p<0.01) (Table3). Glucocorticoid was recommended for the treatment of moderate/severe PCP as the guideline . In this study, glucocorticoid usage rate in moderate/severe group (29.76%) was significantly higher than that in mild group (16.42%, p<0.01) (Table3).
4. Side-effects of trimethoprim/sulfamethozole administration
Although sulfonamide has a good anti-pneumocystis effect, the facts cannot be denied that the high incidence of intolerable negative reaction occurs during standard dose of trimethoprim/sulfamethozole usage which usually interrupts anti-pneumocystis treatment. In this study, the incidence of TMP-SMZ induced epispasis was 6.34% and mild group was higher than moderate/severe group (8.96% vs. 2.93%, p<0.01). TMP-SMZ induced fever was 2.99% and liver injury was 3.38%. TMP-SMZ induced renal injury was rare. TMP-SMZ associated leukopenia was 7.82% and thrombocytopenia was 2.33% and anemia was 2.96%. Severe alimentary tract indisposition was 1.90% (Table4). Totally, the incidence of TMP-SMZ associated side-effects was low and mainly focused on epispasis, fever, liver injury and myelosuppression. the incidence of TMP-SMZ associated side-effects was low and mainly focused on epispasis, fever, liver injury and myelosuppression.
5. Multivariate analysis for predictors of favorable and unfavorable outcome among pneumocystis pneumonia patients
Whether death or not of these patients was selected as the dependent variable, multivariate logistic regression analysis was performed with the variables of demographics, clinical features listed in table 1 and 2 as independent variables to explore the predictors of favorable and unfavorable treatment outcome among PCP. In the multivariate analysis, we found CD4 cell counts were favorable predictor of PCP outcome (OR 0.900, 95% CI 0.813-0.996, P=0.041), and lactate dehydrogenase (OR 1.020, 95% CI 1.006-1.033, P=0.005), alveolar-arterial O2 difference (OR 1.051, 95% CI 1.005-1.099, P=0.030) and neutrophils counts (OR 1.436, 95% CI 1.002-2.060, P=0.049) were unfavorable predictors of PCP outcome (Table5). Further, we constructed a logistic prognostic model [prognostic index=-11.953+0.049×alveolar-arterial O2 difference+0.019×lactate dehydrogenase level+0.362×neutrophils counts-0.106×CD4 cell counts]. A receiver operating characteristic (ROC) curve was employed to identify the prognostic model and area under curve (AUC) was 0.959 (95%CI 0.914-1.000, cut-off value 0.442, P=0.000), with sensitivity and specificity predictive values of 87.5% and 3.6%, respectively (Figure2).