Introduction

Hematological disorders are increasingly being treated over the past two decades in developing countries. Advances in the therapeutic modalities have considerably improved the prognosis in these patients and complete cure can be achieved in a considerable number of patients with hematological disorders₁,₂,₃. However these patients are highly susceptible to overwhelming infections particularly during the period of severe neutropenia. Haemorrhagic complications secondary to severe thrombocytopenia are also common in these patients. It is inevitable that some of these complications may require a multidisciplinary approach and may necessitate intensive care management (ICU), particularly the need for mechanical ventilation and haemodynamic support. The outcome of patients with hematological malignancies requiring ICU care are known to be very poor, but in some patients intensive care management can be life saving₄,₅,₆,₇,₈,₉,₁₀,₁₁,₁₂. This study was conduced with a view to assess the utility and outcome of patients with hematological disorders requiring ICU in a tertiary care center.

Methodology

This retrospective study was conducted at the Christian Medical College & Hospital, a tertiary care 1800-bed university teaching hospital in south India. By reviewing admission records, we identified all patients who were admitted to the Medical intensive care unit with various hematological disorders between January 1998 through December 2001. Information was collected by reviewing the medical records on the demographic characteristics, underlying hematological conditions, indications for ICU admission and duration of ICU stay. The need for various supportive measures (mechanical ventilatory support, hemodynamic support and various other therapeutic interventions) were ascertained. The SOFA (sepsis-related organ failure assessment scoring system) was used to determine the organ dysfunction in these patients during the first 24 to 48 hours of ICU admission. In addition to the laboratory parameters required for the SOFA scoring, other laboratory results, arterial blood gas values, results of culture, biopsy specimens and autopsy reports were obtained when available.

Patients discharged from ICU and transferred back to the ward, regardless of subsequent outcome, were considered as ICU survivors. Patients who were discharged from ICU and subsequently discharged from the hospital were considered as Hospital survivors.

Statistical analysis: Fisher exact test was used when the expected cell value was under 5, otherwise Chisquare test was used to find the statistical significance. A p value < 0.05 was considered to be significant.

Results

There were a total of 3070 hospital admissions with various hematological disorders during the period January 1998 to December 2001. One hundred and four (3.3%) of these patients required admission to the Medical ICU. During the same period there were a total of 3130 admissions to the Medical ICU and 104 (3.3%) of the ICU admissions were due to hematological disorders.

Demographic characteristics and underlying hematological conditions

There were 67 (64%) male subjects and 37 (36%) female subjects with a mean age of 36.3 ± 15.3 years (range 10 to 65 years). The various hematological diagnosis of the study subjects are given in Table 1. Four patients had undergone bone marrow transplantation in the past, two of them were for chronic myeloid leukemia and two for non-Hodgkins lymphoma. Among the 104 patients 2 were tested positive for HIV Elisa and 5 were HbsAg positive. Sixty (58%) patients had received cytotoxic drugs during the current hospital admission or in the immediate past.

Figure 1

Table 1: Hematological diagnosis of the study subjects n = 104.

ICU admission

Investigations and SOFA score

The initial mean PaO₂ was 156 mm Hg ± 104. Forty-five patients had leukopenia with a total leukocyte count less than 3000/cu mm. The SOFA score for organ dysfunction is given in Table 2. The exact Glasgow coma scale (GCS) at admission was difficult to establish accurately in view of the use of sedation in intubated patients, hence GCS was not included in the assessment of neurological status in the SOFA scoring. Of the available data, 18 (17%) had score above 10. The base line investigations other than those included in the SOFA score are given in Table 3.

Figure 2

Table 2: SOFA score for organ dysfunction

Figure 3

Table 3: Investigations other than mentioned in the SOFA scoring

ICU supportive measures

Outcome in ICU and hospital

Diagnosis of condition requiring ICU admission

The final diagnosis was established by utilizing the results of culture, biopsy specimens, autopsy study and relevant noninvasive procedures where applicable. Sepsis was diagnosed in 46 (44%) patients, of them 33 had evidence of bacteremia with the majority being due to gram negative organisms. Pneumonia was diagnosed in 21 (20%) patients and cerebrovascular events in 20 (19%) patients. Table 4 gives the diagnosis in the 104 patients requiring admission to ICU.

Figure 4

Table 4: Conditions requiring admission to medical ICU (N = 104)

Sepsis including bacterial and fungal*

Autopsy findings

Prognostic factors

The prognostic factors and the final outcome were assessed comparing patients who survived till hospital discharge (20) against those who did not survive till hospital discharge (84). All the patients admitted to the ICU following a cardio-pulmonary arrest and those with advanced malignancy did not survive till hospital discharge. Other prognostic factors are given in Table 5.

Figure 5

Table 5: Prognostic factors

Sepsis including bacterial and fungal*

The various ICU supportive measures in combination with lab parameters and organ dysfunction contributing to mortality in the non-survivors are given in Fig 1.

Figure 6

Figure 1: Comparison of various ICU supportive measures and organ dysfunction in non survivors

ANC = Absolute neutrophil count

Discussion

There are published reports on the utility of ICU care and its outcome among patients with hematological disorders from the West₄,₅,₆,₇,₈,₉,₁₀,₁₁,₁₂. However there is paucity of data from the developing countries regarding this aspect. In this study about 3% of the patients with hematological disorders have required admission to Medical ICU, of which 66% had hematological malignancies. The percentage of patients who were bone marrow transplant recipients were few unlike reports from the developed countries₄,₁₃,₁₄. This may be because post bone marrow transplant recipients patients are generally managed in the transplant unit in the absence of the requirement for mechanical ventilation.

The most common reason for ICU admission was respiratory distress or respiratory failure. However many patients had a combination of hemodynamic instability (36%) or low GCS (21%) along with respiratory failure. Previous studies have shown similar results among patients who require ICU care with hematological disorders ₄,₆,₁₀,₁₃,₁₅. Few studies have shown that ICU mortality is best predicted by the diagnosis of respiratory failure at admission₄,₆,₈,₁₁. In one of the above studies₆, only 9 out of 50 patients (18%) admitted with respiratory failure survived to hospital discharge.

Hospital mortality rates have been reported to be high ranging from 69 – 80% when patients with haematological malignancies develop an acute illness of sufficient severity to warrant ICU admission₇. It is noted that the expected mortality among patients with hematological disorders who require ICU care is considerably higher than the predicted mortality compared to those patients admitted with other diseases₄,₆,₂₃. The survival to hospital discharge in this present study was 19% almost comparable to the reports from various centers from the developed countries ranging from 18 – 23%₄,₅,₆,₇,₁₂. Tremblay et al observed less than 3% ICU survival among leukemia patients who required ICU admission₂₃. The survival to hospital discharge among patients with non-malignant conditions was 10% higher than in those with malignant conditions, however this did not attain statistical significance. Our results indicate that patients with non-malignant hematological conditions are also at a high risk if their acute illness warrants ICU admission.

In the present study the need for mechanical ventilation was significantly associated with high mortality. Sixty-six out of the 69 patients who required mechanical ventilation did not survive to hospital discharge. Jackson et al ₄ noted only 17% survival among those who required mechanical ventilation while Price et al ₁₇ reported a survival of 19%. Other studies have reported a mortality of 87%₆ and 74%₉. The high mortality noted in patients requiring mechanical ventilatory support only indicates that these patients have far advanced and serious underlying conditions. However the trend towards survival in mechanically ventilated patients is noted to be increasing₁₈,₁₉.

Cerebro-vascular event also contributed to significant number of ICU admission in this study. Majority of these were due to intracranial bleeds secondary to severe thrombocytopenia. Forty-eight patients had a platelet count less than 20000/mm³. Most patients admitted with cerebro-vascular event did not survive to hospital discharge, but this value did not attain statistical significance. Earlier reports had shown that thrombocytopenic patients in ICU generally have a poorer prognosis compared to non-thrombocytopenic patients₂₀. Mortality in ICU among patients with severe neutropenia is noted to be high₁₁. In the present study 67% of the patients had drug induced cytopenia and the mortality was significantly associated with an ANC count less than 500/mm³.

Earlier published reports have shown that the requirement of ICU supportive measures for organ dysfunction is significantly associated with mortality among patients with hematological disorders who require ICU care₈,₁₀,₁₁. Available data from this present study show that the number of patients with significant organ dysfunction is less among the survivors compared to non-survivors, indicating poorer prognosis for those who are admitted with significant organ dysfunction. Studies from our center have shown that infection with Gram negative organisms and fungal infections are the most common among patients with hematological malignancy₂₁,₂₂,₂₃ and in post transplant patients₂₄. In this present study, presence of sepsis was associated with a high mortality. Positive microbiological identification of an organism did not improve the outcome compared with those patients who were diagnosed to have probable sepsis. A combination of mechanical ventilation, hemodynamic support, ANC count < 500/mm³ and platelet counts < 20000/mm³ were noted to be predictors of high mortality in those who did not survive to hospital discharge. The SOFA scoring system is shown to be useful in assessing organ dysfunction in ICU and studies have shown higher mortality in patients with higher scores₁₅,₂₅. We could not assess the neurological dysfunction by using this scoring system retrospectively. Our results indicate that SOFA score can be used with ease to assess the organ dysfunction, but the neurological dysfunction may be difficult to assess in view of sedation used in these patients.

We conclude that patients admitted following a cardio-pulmonary arrest, severe neutropenia with an ANC count < than 500/mm³, severe thrombocytopenia with a platelet counts < 20,000/mm³, patients requiring mechanical ventilatory and hemodynamic support have a poor prognosis in the ICU. The addition of other organ dysfunction further increases this mortality. The utility of non invasive ventilation among patients with hematological disorders need to be evaluated, to see if it could reduce the mortality as reported from other centers₂₆,₂₇,₂₈.

Correspondence to

Dr. Biju George Lecturer Dept of Haematology CMC Hospital Vellore – 632004 Tamil Nadu India Ph No: +91 – 0416 – 2222102 Ex 2352 Fax No: +91 – 0416 – 2232035/2232054 e-mail: biju@cmcvellore.ac.in