Rheumatology International

, 29:491

Risk factors of Pneumocystis jeroveci pneumonia in patients with systemic lupus erythematosus


    • Division of Allergy Immunology and Rheumatology, Department of Medicine, Faculty of Medicine at Ramathibodi HospitalMahidol University
  • Kitti Totemchokchyakarn
    • Division of Allergy Immunology and Rheumatology, Department of Medicine, Faculty of Medicine at Ramathibodi HospitalMahidol University
  • Kanokrat Nantiruj
    • Division of Allergy Immunology and Rheumatology, Department of Medicine, Faculty of Medicine at Ramathibodi HospitalMahidol University
  • Suchela Janwityanujit
    • Division of Allergy Immunology and Rheumatology, Department of Medicine, Faculty of Medicine at Ramathibodi HospitalMahidol University
Original Article

DOI: 10.1007/s00296-008-0721-6

Cite this article as:
Lertnawapan, R., Totemchokchyakarn, K., Nantiruj, K. et al. Rheumatol Int (2009) 29: 491. doi:10.1007/s00296-008-0721-6


Pneumocystis jeroveci pneumonia (PCP) is an opportunistic infection which occurs mostly in the immune-deficiency host. Although PCP infected systemic lupus erythematosus (SLE) patient carries poor outcome, no standard guideline for prevention has been established. The aim of our study is to identify the risk factors which will indicate the PCP prophylaxis in SLE. This is a case control study. A search of Ramathibodi hospital’s medical records between January 1994 and March 2004, demonstrates 15 cases of SLE with PCP infection. Clinical and laboratory data of these patients were compared to those of 60 matched patients suffering from SLE but no PCP infection. Compared to SLE without PCP, those with PCP infection have significantly higher activity index by MEX-SLEDAI (13.6 ± 5.83 vs. 6.73 ± 3.22) or more renal involvement (86 vs. 11.6%, P < 0.01), higher mean cumulative dose of steroid (49 ± 29 vs. 20 ± 8 mg/d, P < 0.01), but lower lymphocyte count (520 ± 226 vs. 1420 ± 382 cells/mm3, P < 0.01). Interestingly, in all cases, a marked reduction in lymphocyte count (710 ± 377 cells/mm3) is observed before the onset of PCP infection. The estimated CD4+ count is also found to be lower in the PCP group (156 ± 5 vs. 276 ± 8 cells/mm3). Our study revealed that PCP infected SLE patients had higher disease activity, higher dose of prednisolone treatment, more likelihood of renal involvement, and lower lymphocyte count as well as lower CD4+ count than those with no PCP infection. These data should be helpful in selecting SLE patients who need PCP prophylaxis.


Pneumocystis jeroveci pneumonia (PCP), previously known as Pneumocystis carinii [1] is an opportunistic infection which occurs almost exclusively in the immunocompromised host with and without acquired immune deficiency syndrome (AIDS) [2, 3]. The incidence of PCP was increase during the AIDS era with a marked decline in mid 1992 probably due to the introduction of PCP prophylaxis [4]. On the contrary, PCP in non-AIDS connective tissue disease patients was increased [517] and showed higher mortality rate than PCP in AIDS patients [2, 5]. Systemic lupus erythematosus (SLE) with PCP was also observed to be acute at onset, severe and most of all with high mortality [2]. Previous studies [2, 14, 18, 19] indicated the risk factors for PCP infection in SLE as lymphopenia, increased LDH, interstitial fibrosis, high dose of corticosteroid and cytotoxic drugs. The aim of our study is to further identify the risk factors of PCP infection in Thai patients with SLE which may add on useful information for establishment of guideline for PCP prophylaxis in SLE.

Patients and methods

Source of data

Patients were identified through a search of Ramathibodi hospital medical records between January 1994 and March 2004.

Diagnostic criteria

Fifteen cases of SLE with PCP were identified by ICD-10. The diagnosis of SLE fulfilled the 1982 American College of Rheumatology revised classification criteria for SLE, while the diagnosis of PCP was made by (1) a compatible clinical presentation including dyspnea, fever, cough, and in most cases a newly recognized infiltrate on the chest roentgenogram; (2) an associated underlying condition compromising immunologic function; and (3) sputum or branchoalveolar lavage organisms that was dectected by using Grocott Methenamine Silver stain or immunofluorescence antibody staining (IFA) [3].

Research method

The 60 patients without PCP in the control group were selected by the table of randomization from the consecutive age and sex matched SLE patients who were monitored regularly over the same period at Ramathibodi hospital. The usual clinical and biological data at diagnosis and during the follow-up period were collected. PCP infection usually occurs within 6–7 months after initiation of SLE immunosuppression therapy. The control group is also selected from the same period (after treatment) for comparison. The doses and mode of administration of corticosteroid and cytotoxic drugs were recorded. Any subject with the human immunodeficiency virus (HIV) was excluded from the study.

Statistical analysis

Data from patients and controls were compared using the Chi-square test for categorical variables, and the Mann–Whitney U test for quantitative variables. The hypothesis was examined based on two-tailed P values. The test values of 0.01 or less were considered as statistically significant. We also used the Odd ratio to determine the significant of the risk factor for PCP.


From retrospective analysis of the medical records during January 1994–March 2004, we found 15 cases of PCP infected SLE patients. Three patients died, making the mortality rate of 20%. The PCP infected SLE group had higher Mexican Systemic Lupus Erythematosus (MEX-SLEDAI) [20] than the control group (13.6 ± 5.83 vs. 6.73 ± 3.22, P < 0.01). Renal involvement was also more common in the PCP group (86 vs. 11.6%, P < 0.01) with the Odd ratio of 49.214 (95% CI 38.76–54.42) as shown in Table 1 and Fig. 3. The incidence of interstitial pulmonary fibrosis (IPF) was not significantly different between the two groups (13.33 vs 6.66%, P < 0.01). There was also no statistically significant difference in the incidence of cutaneous involvement, oral ulcer, arthritis, serositis, hematologic involvement, CNS involvement, and antiphospholipid syndrome (Table 1; Fig. 3). As for treatment comparing the 6-month mean dose before infection, the PCP group took higher prednisolone dose (49 ± 29 vs. 20 ± 8 mg/d, P < 0.01; Odd ratio 7.81, 95% CI 4.71–12.06) but similar cyclophosphamide (78 ± 33 vs. 74 ± 18 mg/d), azathioprine (62 ± 50 vs. 54 ± 20 mg/d) and chloroquine (230 ± 8 vs. 238 ± 10 mg/d) doses. Among the laboratory values recorded during the study, the only parameter that registered a significant variation between the two groups was the lymphocyte count. The mean lymphocyte count was calculated from the amount of lymphocyte during the 6–7 months after initiation of treatment for SLE in control which was the comparable period as 6–7 months before the occurrence of PCP in infected group. Although both SLE groups had lymphocytopenia, the lowest lymphocyte count was significantly less in the PCP group (520 ± 226 vs. 1420 ± 382 cells/mm3, P < 0.01; Odd ratio 18.97, 95% CI 6.77–22.4). This significant statistical difference was also present in the mean lymphocyte count (1040 ± 438 cells/mm3 vs. 1,842 ± 632 cells/mm3, P < 0.01; Odd ratio 10.20, 95% CI 5.69–18.21). Interestingly, the lymphocyte count has declined to the lowest level before the onset of PCP (mean 710 ± 377 cells/mm3; Fig. 1). This declination is determined by the difference between the mean value and the lymphocyte count immediately before the onset of PCP (329 ± 210 cells/mm3). The decrement in lymphocyte count is strong enough to be used as a critical parameter of developing PCP infection in these immunocompromised patients (Fig. 2).
Table 1

Demographic and biological data of SLE patients with and without PCP


With PCP

Without PCP

P value

Number of case



NS (>0.01)

Sex (F:M)




Age (mean ± SD)

37 ± 10

37.1 ± 14.3


Organ involvement


13.6 ± 5.83

6.73 ± 3.22


Cutaneous involvement

2 (13.33%)

7 (11.66%)


Oral ulcer

3 (20%)

16 (26.66%)



9 (60%)

39 (65%)



1 (6.67%)

3 (5%)


Renal involvement

13 (86%)

7 (11.66%)


Hematologic involvement

Autoimmune hemolytic anemia



11 (73%)

11 (73%)

5 (33.33%)

23 (30.33%)

37 (61.66%)

12 (20%)




Central nervous system involvement

3 (20%)

16 (26.66%)


Interstitial pulmonary fibrosis

2 (13.33%)

4 (6.66%)


Antiphospholipid syndrome

2 (13.33%)

5 (8.33%)


Prednisolone (mean ± SD)

49 ± 29

20 ± 8


Cyclophosphamide (mean ± SD)

78 ± 33

74 ± 18


Azathioprine (mean ± SD)

62 ± 50

54 ± 20


Chloroquine (mean ± SD)

230 ± 8

238 ± 10


White blood cell counts (mean ± SD)

6,462 ± 1943

7,700 ± 2100


Polymorphoneuclear cell counts (mean ± SD)

4,995 ± 1642

5,870 ± 780


Lymphocyte counts (mean ± SD)

1,040 ± 438

1,842 ± 632


Minimal lymphocyte counts (mean ± SD)

520 ± 226

1,420 ± 382


Lymphocyte counts when PCP occurs

710 ± 377



Delta lymphocyte (mean lymphocyte counts − lymphocyte counts at onset PCP)

329 + 210



Estimated CD4

156 ± 5

276 ± 8


Mann–Whitney U test P = 0.01 is significant difference, Chi-square P = 0.01 significant difference

NS  non statistic significant

Fig. 1

Mean Lymphocyte in PCP and control at the comparable time period after initiation of immunosuppression treatment until before PCP infection

Fig. 2

The declination of the lymphocyte count prior to the onset of PCP infection in SLE patients

Fig. 3

Odd ratio of various risk factors

Although we did not measure CD4 count, the estimated CD4 count, which can be calculated from 15–20% of lymphocytes, was lower in PCP infected patients than in the controls (156 ± 5 vs. 276 ± 8; Odd ratio 6.71, 95% CI 4.42–9.46).


From our study, the mortality rate of SLE patients infected by PCP is 20%. This high mortality rate signifies the importance of PCP prophylaxis in high risk SLE patients as has been done in AIDS patients [21].

Upon exploring the risk factors for developing PCP, we have found that the mean lymphocyte counts and the lymphocyte counts during treatment for SLE in the PCP group were significantly lesser than in the control group. The mean lymphocyte count at 1,000 cells/mm3 and minimal lymphocyte count at 750 cells/mm3 after treatment for SLE defined the groups of patients at greater risk of PCP infection (Table 1). Moreover, a decrease in lymphocyte count was always found prior to the onset of PCP infection (Fig. 2). This finding was in agreement with the studies by Porges et al. [2], Kadoya et al. [14] and Godeau et al. [12]. They also indicated that the decrease in lymphocyte count can be used as a predictor for PCP infection in SLE patients (Table 2). However, as lymphocytopenia is a marker for active SLE as well, we would recommend that the profound decrease in lymphocyte count at or below 750 cells/mm3 at any time during treatment signifies the necessity for primary PCP prophylaxis. The guideline for giving the prophylactic treatment in AIDS patients whose CD4 count were <200 [23] can also be applied to SLE patients as the estimated CD4 count (mean ± SD) in our PCP infected group was 156 ± 5 compared to 276 ± 8 in the control group.
Table 2

Comparison of risk factors and suggestions for PCP prophylaxis between various studies


Godeau et al. [12]

Li et al. [25]

Kadoya et al. [14]

Galeazzi et al. [22]

Porge et al. [2]

Lertnawapan et al.


Wegener’s granulomatosis

Connective tissue disease





Mean/median lymphocyte count in PCP versus control (cells/mm3)

preRx. 1,060 vs. 1,426

352 ± 152 in PCP

1052.7 ± 309.1 vs. 1841.6 ± 632.4


595 vs. 833

1,040 ± 438 vs. 1,842 ± 632

Lowest lymphocyte count in PCP versus control (cells/mm3)

244 versus 738

959.3 ± 196 vs. 1663.9 ± 682.4

520 ± 226 vs. 1420 ± 382

Lymphocyte at onset PCP (cells/mm3)


710 ± 377

CD4 (cells/mm3)


87 ± 78


156 ± 5

Prednisolone or immunosuppressive treatment

Cyclophosphamide 1.67 versus 0.97 mg/kg per day

Prednisolone 25–60 mg/day and Immunosuppressive

No significant of dose of Prednisolone

Prednilosone 43 vs. 20 mg/day

Prednisolone 49 ± 29 vs. 20 ± 8 mg/day




 CD4 < 250

 High dose immunosuppressive


 Lymphocyte < 600

 CD4 < 250

 Cytotoxic treatment

Risk of PCP

 Lower Lymphocyte

 Interstitial  Pulmonary Fibrosis


 Low CD4


 Lymphocyte < 350

 High steroid dose


 Lymphocyte < 750

 Prednisolone > 20 mg/day

 CD4 < 150

 Renal involvement or hight disease activity

Other significant risk factors for developing PCP demonstrated by our study were high disease activity, renal involvement and high prednisolone dosage. The result that MEX-SLEDAI was higher in the PCP group than in control confirmed the effect of the severity of disease on the occurrence of PCP.

Contrary to the study by Kadoya et al. [14] which demonstrated interstitial pulmonary fibrosis as a significant risk factor, our study showed that only the renal involvement correlated significantly with the higher incidence of PCP infection.

All our 15 PCP infected SLE patients had been treated with prednisolone at the mean dosage of 49 ± 29 mg/day which was significantly higher than the mean dose in the control group. This is in contrast with previous study by Kadoya et al. [14], which showed no significant correlation between the steroid dosage and the risk of PCP infection. Nevertheless, our finding is in agreement with the recommendation to start the PCP prophylaxis in patients who received more than 20 mg prednisolone equivalent per day [24].

Unlike previous study by Godeau et al. [12], we could not find any significant correlation between PCP infection and the use of other immunosuppressive agents such as cyclophosphamide and azathioprine (Table 2).

Although this study is limited by the small number of PCP infected SLE patients and insufficient laboratory data for analysis, a demonstration of a profound decrease in lymphocyte count prior to the onset of PCP infection should be clear enough to be included in the standard guideline for PCP prophylaxis in SLE.

Copyright information

© Springer-Verlag 2008