- Top of page
- Patients and methods
Procalcitonin (PCT) has proven to be a very sensitive marker of sepsis for non-leucopenic patients. Little is known about its relevance in immunosuppressed and leucopenic adults. Four hundred and seventy-five PCT determinations were carried out in 73 haematological patients: on 221 occasions the white blood cell (WBC) count was < 1·0 × 109/l and on 239 occasions it was > 1·0 × 109/l leucocytes. Patients were classified as: non-systemic infected controls (n = 280), patients with bacteraemia (n = 32), sepsis (n = 30), severe sepsis (n = 3), septic shock (n = 3) and systemic inflammatory response syndrome (SIRS) (n = 62). When the WBC count was > 1·0 × 109/l, gram-negative bacteria induced higher PCT levels (median 9·4 ng/ml) than gram-positives (median 1·4 ng/ml). In cases with a WBC < 1·0 × 109/l, PCT levels were similar for gram-negative and gram-positive bacteria (1·1 ng/ml versus 0·85 ng/ml). Regardless of the leucocyte count, the median PCT level in bacteraemia cases always remained < 0·5 ng/ml. In heavily leucopenic situations, PCT levels were never > 2 ng/ml even in the sepsis and severe sepsis/septic shock groups, whereas a WBC count > 1·0 × 109/l resulted in median PCT values of 4·1 ng/ml and 45 ng/ml respectively. The positive predictive value for sepsis (cut-off 2 ng/ml) was 93% in cases of WBC count > 1·0 × 109/l, but only 66% in leucopenic conditions. The negative predictive value (cut-off 0·5 ng/ml) was 90% when the WBC count was > 1·0 × 109/l and 63% in leucopenic conditions. Procalcitonin is an excellent sepsis marker with a high positive- and negative-predictive value in patients with WBC count > 1·0 × 109/l, but it does not work satisfactorily below this leucocyte count.
Sepsis, severe sepsis, septic shock and sepsis induced multi-organ failure are the most important causes of death in immunosuppressed neutropenic patients. Furthermore, because of the disease- and treatment-associated immunosuppression of these patients, diagnosis of sepsis is often difficult. When the correct diagnosis of sepsis is delayed, there is a high mortality rate, and this is the reason why a highly specific, rapid and cost-efficient marker of sepsis is warranted. In several studies, C reactive protein (CRP), tumour necrosis factor-α (TNF-α), interleukin 6 (IL-6) and interleukin 8 (IL-8) have been reported to be increased in viral and/or bacterial infections (Bruserud et al, 1994; Manian, 1995; Engel et al, 1998). Procalcitonin (PCT), a precursor protein of calcitonin, is a specific and sensitive marker of severe bacterial infections in non-neutropenic patients (Chiesa et al, 1998; Oberhoffer et al, 1999). For example, PCT was shown to be a useful marker in differentiating viral from bacterial meningitis in children (Gendrel et al, 1999) and also proved to be discriminant in this differential diagnosis in a study of 179 consecutive adult patients with suspected meningitis (Viallon et al, 2000). Furthermore, it has been demonstrated that a persistent or an increasing PCT value is associated with a high mortality rate in children with septic shock (Hatherill et al, 2000). Only few data are available regarding PCT as an inflammation marker in neutropenic patients (Engel et al, 1999; Fleischhack et al, 2000).
The source of PCT production during infection is not known. PCT levels are increased particularly in patients with small cell carcinoma of the lung (Cate et al, 1986). Another observation suggested that the neuroendocrine cells of the lungs may be a site of PCT production (Maruna et al, 2000). Results of a recent study indicate that the hepatocytes are likely to be the primary PCT producers (Nijsten et al, 2000) and finally, a possible role for peripheral blood mononuclear cells as a PCT source has been discussed (Oberhoffer et al, 1999). Because of this we have investigated whether PCT had a good sensitivity and specificity in heavily leucopenic and immunosuppressed patients.
Patients and methods
- Top of page
- Patients and methods
Between March 1999 and August 2000, samples were taken from 73 patients at the time of admission to the haematological department of the Regional Hospital Bozen for various reasons, such as initiation of chemotherapy and fever with or without neutropenia. Axillary temperature, systolic and diastolic blood pressure, heart rate, respiratory frequency, urinary flow as well as the white blood cell (WBC) count and PCT were determined. After analysis of all diagnostic criteria, patients were classified according to the criteria of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee (American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference, 1992). Although a bacteraemia was defined by the presence of viable bacteria in the blood, definitions of sepsis (n = 30), severe sepsis (n = 3) and septic shock (n = 3) required a positive-blood culture or broncho-alveolar lavage in combination with two or more of the following conditions: temperature > 38°C or < 36°C; heart rate > 90 beats/min; paO2 < 72 mmHg. An event was defined as systemic inflammatory response syndrome (SIRS n = 62) when two or more of the following conditions were present: temperature > 38°C or < 36°C; heart rate > 90 beats/min; paO2 < 72 mmHg and negative haemo- or uroculture.
All PCT and haemoculture samples were taken at first appearance of fever (defined as fever ≥ 38°C). The control group consisted of 280 events in afebrile patients and patients with localized infection but no laboratory, haemodynamic or clinical signs of sepsis. In case of fever, a careful physical examination was carried out and targeted samples were taken. As routine in haematological units, oropharyngeal, anal, genital swabs, smears from dermatological lesions and urocultures were taken. A radiography of the chest was carried out in every case. If the axillary temperature was ≥ 38°C, an aerobic and an anaerobic blood culture were taken and microbiological analyses carried out according to standard procedures.
Procalcitonin was determined by a commercially available immunoluminetric assay LUMI Test PCT kit (B.R.A.H.M.S. Diagnostica, Berlin Germany). Values > 0·5 ng/ml were considered to be pathological in accordance with the manufacturer's instructions (Al-Nawas & Shah, 1996). Leucocyte counts were measured in the hospital's haematological laboratory using a MaxM (Coulter, IL Milan Italy).
Statistical analysis The statistical significance was calculated using the Mann–Witney U-test and probabilities of < 0·05 were accepted as statistically significant.
- Top of page
- Patients and methods
Four hundred and seventy-five PCT values in 73 patients with different haematological pathologies were determined (Table 1A). Eighty-three (17·4%) levels were > 2 ng/ml, 128 (26·9%) were between 0·5 and 2 ng/ml and 264 (55·5%) were < 0·5 ng/ml. Patients affected by acute myeloid leukaemia (AML), acute lymphoblastic leukaemia (ALL) and Non Hodgkin's lymphoma (NHL) showed similar percentages of PCT values > 2 ng/ml, whereas multiple myeloma patients had significantly more PCT-determinations > 2 ng/ml (P < 0·05).
Table 1. (A) Patients characteristics.
| || ||Sex||PCT (ng/ml) |
| Pathology|| Total|| Male|| Female|| Total||> 0·5 (%)||0·5–2 (%)||> 2 (%)|
|AML||33||17||16||286||166 (85)||69 (24)||51 (17)|
|ALL||5||3||2||40||25 (62)||12 (30)||3 (7)|
|NHL||20||7||13||82||53 (64)||22 (26)||7 (8)|
|MM||12||7||5||48||20 (41)||13 (27)||15 (31)|
|CMML||2||2||0||18||0||12 (66)||6 (33)|
Leucocyte counts were available for 460 PCT measurements: 221 were carried out for leucopenia, 239 when more than 1·0 × 109/l leucocytes were measured (P = NS). When the leucocyte count was < 1·0 × 109/l, 37% of all PCT determinations were > 0·5 ng/ml compared with 52% in situations with more than 1·0 × 109/l leucocytes (P = 0004) (Table 1B).
Table 2. (B) Comparison of PCT values in leucopenic and non-leucopenic instances.
| ||PCT (ng/ml) |
|Leucocytes||< 0·5 (%)||0·5–2 (%)||> 2 (%)|
|< 1·0 × 109/l||140 (63)||51 (23)||30 (14)|
|> 1·0 × 109/l||114 (48)||75 (31)||50 (21)|
Swabs were taken during neutropenia and in the presence of fever. Altogether, 44 swabs were positive for gram-positive bacteria, 32 for fungi, 30 for gram-negative bacteria, 27 were negative and on one occasion, mycobacterium tuberculosis was detected. Of 93 routinely taken urocultures, 91 (97·8%) were negative and two cases (2·1%) of gram-negative bacteria were identified. In six bronchoalveolar lavages, three mycoplasma infections and one Chlamydia infection were detected, whereas two procedures gave no positive result.
A total of 287 haemocultures were taken: 42/287 (14·6%) blood samples were positive for gram-positive bacteria, 26/287 (9%) for gram-negative bacteria, 7 (2·4%) showed fungi and 2 (0·6%) viruses; 210 (73·1%) samples were negative. A comparison of positive-haemoculture results in leucopenic and non-leucopenic patients is shown in Table II. In 26 blood cultures, a gram-negative bacterium was identified. Similar numbers of fungi and gram-negative bacteria were identified in leucopenic and non-leucopenic patients, whereas gram-positive bacteria were identified more often in cases with a WBC count < 1·0 × 109/l. As shown in Table III, similar numbers of bacteraemia, sepsis and severe sepsis/septic shock caused by gram-negative bacteria were present in 13 leucopenic and 13 non-leucopenic events. A gram-positive bacterium could be identified in 42 blood cultures. In leucopenic situations, more gram-positive blood cultures were found; from the clinical point of view, the majority of those clinical situations were defined as bacteraemia. No severe sepsis or septic shock caused by gram-positive bacteria was seen.
Table II. Pathogens detected in haemocultures during leucopenic and non-leucopenic situations.
|Haemocultures||WBC < 1·0 × 109/l||WBC > 1·0 × 109/l|
Table III. Comparison of syndromes caused by gram-negative and gram-positive bacteria in leucopenic versus non-leucopenic events.
|WBC < 1·0 × 109/l||WBC > 1·0 × 109/l||Total||WBC < 1·0 × 109/l||WBC > 1·0 × 109/l||Total|
The median PCT level in the presence of gram-negative bacteria was 4·1 ng/ml (range 0·3–44·9 ng/ml) versus 1·4 ng/ml (range 0·1–61·8 ng/ml) if gram-positive bacteria were identified in the haemoculture (P = < 0·05). A remarkable, although not statistically significant difference for PCT values was found with regard to the identification of gram-negative bacteria in the presence of more or less than 1·0 × 109/l of leucocytes: the median level of PCT was 1·1 ng/ml (range 0·3–14·4 ng/ml) if the WBC count was < 1·0 × 109/l, and 9·4 ng/ml (range 0·4–44·9 ng/ml) when leucocytes > 1·0 × 109/l. No significant difference was evident if a gram-positive bacterium was identified: the median PCT value was 0·85 ng/ml (range 0·1–11 ng/ml) for a WBC count < 1·0 × 109/l, and 1·4 ng/ml (range 0·6–9 ng/ml) for non-leucopenic cases.
Median PCT levels were 0·3 ng/ml (range 0·1–19·6 ng/ml) for bacteraemia, 1·65 (0·1–61·8 ng/ml) for sepsis, 33·1 ng/ml (range 0·3–56·9 ng/ml) in the presence of severe sepsis and septic shock and 0·3 (range 0·1–31 ng/ml) for patients with localized infection but no laboratory, haemodynamic or clinical signs of sepsis.
A total of 44 PCT determinations was undertaken in instances with positive haemoculture and a WBC count < 1·0 × 109/l. In these situations, median PCT levels did not exceed 2 ng/ml. In contrast, 25 PCT determinations were available in the presence of a positive blood culture and a WBC count > 1·0 × 109/l. The median PCT levels were 0·2 ng/ml for bacteraemia, 4·1 ng/ml for sepsis and 45 ng/ml for severe sepsis/septic shock.
The sensitivity of the test for sepsis, severe sepsis and septic shock was 63% when the WBC count was < 1·0 × 109/l and 94% when > 1·0 × 109/l leucocytes were counted. The specificity was 60% in both cases.
A difference in the positive prognostic value for sepsis, severe sepsis and septic shock was found when comparing events with more or less than 1·0 × 109/l WBCs: at a cut-off of 2 ng/ml the positive prognostic value was 66% for a WBC count < 1·0 × 109/l, whereas it was 93% when > 1·0 × 109/l leucocytes were counted. On the other hand, the negative prognostic value for PCT levels lower than 0·5 ng/ml was 63% in leucopenic situations, but 90% when the haemochrome presented > 1·0 × 109/l leucocytes.
- Top of page
- Patients and methods
PCT is a sensitive marker for systemic bacterial infections in non-leucopenic patients and serum concentrations seem to correlate with the severity of infection (Assicot et al, 1993). In this study, PCT was shown to demonstrate a clearly inferior value in leucopenic patients. Different sources of this prohormone have already been discussed. (Oberhoffer et al, 1999 ;Maruna et al, 2000; Nijsten et al, 2000). PCT levels in patients with generalized infections can increase up to 10 000-fold from baseline within 24 h (Nylen et al, 1997). Such an amount of PCT cannot be produced solely by neuroendocrine or small cell carcinoma cells (Bertagna et al, 1978), which is the reason why an alternative or primary source of PCT production has to be sought. The most promising candidates for the PCT production could be immunoreactive cells because elevated PCT levels have been found after different stimuli such as heatstroke (Nylen et al, 1997), therapy with OKT3 (Eberhard et al, 1998) and after lipopolysaccharide (LPS) injection in human volunteers (Dandona et al, 1994). More recently, Oberhoffer et al (1999) have demonstrated an intracellular PCT detection in mononuclear cells. On the other hand, some authors, based on reduced PCT production in whole blood samples stimulated with lipopolysaccharides, deny PCT production by circulating blood cells (Monneret et al, 1999).
Another reason why an involvement of leucocytes as a possible PCT source has to be discussed comes from the fact that some authors (Al-Nawas et al, 1996) reported lower PCT levels during leucopenia. This is in agreement with the data found in this study. Because hepatic cells are now also considered to be a source of PCT production (Nijsten et al, 2000), liver damage – for example caused by chemotherapy – could be responsible for a reduced PCT level in haematological patients. Our data does not support this hypothesis because leucopenic and non-leucopenic patients under chemotherapy were investigated in this study and, as shown in Table 1B, PCT values > 0·5 ng/ml were significantly less common in cases of leucopenia.
Further, this study has clearly demonstrated that during sepsis gram-negative bacteria induce significantly higher PCT-levels than gram-positive ones. The LPS structure of the outer membrane of gram-negative bacteria could be responsible for this observation. The endotoxin LPS has different biological properties, such as the activation of macrophages and the subsequent induction of the production of endogenous pyrogenic substances like IL-1 and TNF. For the daily clinical practice this observation has an important consequence. Because the positive predictive value is high, a high PCT level in a non-leucopenic patient is always a sign of serious condition. On the other hand, the PCT determination in heavily leucopenic patients is of limited value because the PCT levels in patients with a WBC count < 1·0 × 109/l are significantly lower in septic conditions even if induced by gram-negative bacteria. A value between 0·5 and 2 ng/ml is an indicator of a serious and life-threatening condition, even if the specificity of the test in heavily leucopenic patients only reaches 60%.
In summary, it can be concluded that PCT is an excellent sepsis marker with a high positive- and negative- predictive value in instances of a WBC count > 1·0 × 109/l, but it is not suitable as a marker below this leucocyte level.