Early Warning Score: A Novel Clinical Score to Stratify Risk in Febrile Thrombocytopenia

Introduction

Febrile thrombocytopenia is a medical condition characterized by the concurrent presence of fever and a low platelet count, often seen in tropical regions with high incidence of infectious diseases. This condition can arise from various infections, such as dengue, malaria, leptospirosis, septicemia, etc, all of which can lead to life-threatening complications if not properly managed.^1-3 Febrile thrombocytopenia is of particular concern due to the increased risk of bleeding, organ damage and mortality. The complexity of managing this condition stems from the varying degrees of thrombocytopenia and the potential for rapid clinical deterioration, necessitating a robust and reliable method of risk stratification to guide clinical interventions.⁴

Platelets play a critical role in maintaining hemostasis, and a significant reduction in platelet count (thrombocytopenia) can lead to bleeding complications, particularly in febrile conditions.^5,6 In febrile thrombocytopenia, platelet counts can drop drastically due to either impaired production or increased destruction, often exacerbated by the body’s immune response to the underlying infection. Platelet count below 150,000/µL is considered thrombocytopenia, with counts under 10,000/µL often associated with severe bleeding risk. However, the correlation between platelet count and bleeding risk is not always straightforward, as some patients with low platelet counts may not exhibit bleeding symptoms, while others with higher counts may experience significant hemorrhages.⁷ This variability has highlighted the need for a more comprehensive approach to patient management.

Historically, platelet transfusions have been administered based primarily on platelet count thresholds. Many clinicians have relied on a count of less than 10,000/ µL as a benchmark for initiating transfusions to prevent spontaneous bleeding, especially in non-bleeding patients.^8,9 However, emerging guidelines, such as those from the Council of Europe, emphasize that transfusion decisions should not be based solely on platelet counts, but should also consider the patient’s overall clinical condition, including the presence of bleeding symptoms and other risk factors. Over-reliance on platelet counts can lead to unnecessary transfusions, which may strain blood banks and expose patients to potential risks such as transfusion reactions or infections.^10,11 Therefore, a more refined and individualized approach to transfusion is necessary to optimize patient care and resource utilization. 

To address the limitations of using platelet count alone for managing febrile thrombocytopenia, Kshirsagar et al., developed a risk assessment scoring system aimed at improving clinical decision-making.^12,13 This scoring system, known as the Early Warning Score (EWS), incorporates multiple clinical parameters such as temperature, respiratory rate, pulse rate, blood pressure, and organ-specific complications (neurological, respiratory, hepatic, hematological and renal) in addition to platelet count. The goal of this system is to provide a more comprehensive assessment of a patient’s condition, allowing for more accurate risk stratification and targeted therapeutic interventions. By considering both vital signs and organ involvement, the EWS helps clinicians identify patients who are at high risk of deterioration and may benefit from more aggressive interventions, including ICU admission and platelet transfusion.^14,15

Despite the promising results of the Kshirsagar scoring system, there is still a need for ongoing research and refinement of this tool to ensure its applicability across diverse patient populations, and hence this study is relevant in this endeavour.^16-18

Materials and Methods

This retrospective study was conducted at the Department of General Medicine, Kempegowda Institute of Medical Sciences, Bengaluru, over a period of six months. The study included 50 patients admitted with febrile thrombocytopenia, defined as fever (temperature >99.9°F) accompanied by a platelet count below 150,000/µL.

Patients included in the study were aged 18 years or older, of either sex, and had a confirmed diagnosis of febrile thrombocytopenia. Patients were excluded if they had fever without thrombocytopenia, or vice versa. Also, patients with primary thrombocytopenic disorders, drug-induced thrombocytopenia, or those undergoing treatment for hematological disorders/malignancies (such as immune thrombocytopenic purpura, or those on chemotherapy/ immunosuppressants), or chronic liver disease were not included.

The medical records of the eligible patients were reviewed, and the following data were extracted: platelet counts, vital signs (pulse rate, systolic blood pressure, respiratory rate, temperature) and complications affecting major organ systems, including the central nervous system (CNS), respiratory, hepatic, renal, and hematological systems. Clinical course and outcomes, including the need for ICU care, bleeding manifestations and mortality, were noted. The EWS, developed by Kshirsagar et al., (Table 1), was applied to each patient based on these parameters, and they were categorized into three risk groups: low (≤7), moderate (8-15), and high (≥16). The study protocol obtained clearance from the institutional ethical committee.

Statistical Analysis

The collected data were analysed using SPSS (Version 22.0). Descriptive statistics were used to summarize the demographic and clinical characteristics. Means and standard deviations were calculated for quantitative variables, while frequencies and percentages were used for categorical variables. Chi-square tests were employed to assess associations between categorical variables, and the Mann-Whitney U test was used to compare continuous variables across risk groups. A P-value less than 0.05 was considered statistically significant.

Results

The study included 50 patients with a mean age of 35.5 years (±13.18). The majority were male (58%), while 42% were female. Hepatic complications were the most common, affecting 92% of the patients, followed by respiratory (32%) and hematological complications (26%). Renal and central nervous system complications were less frequent, observed in 12% and 2% of the patients, respectively. ICU care was required for 18% of patients. Regarding risk classification, most patients (76%) were categorized as moderate risk, while 18% were in the high-risk group, and 6% in the low-risk group.

The study population had a mean platelet count of 74,760/ mm³, mean systolic blood pressure of 109.2 mmHg, mean pulse rate of 94.26/minute, mean respiratory rate of 21.6/minute and mean body temperature of 99.32°F.

The platelet count distribution among the patients revealed that the majority (54%) had platelet counts between 50,000 and 100,000/mm³. Approximately 20% of patients had platelet counts above 100,000/mm³, while 18% fell within the 20,000-50,000/mm³ range. A smaller proportion (8%) had critically low platelet counts below 20,000/mm³. This distribution highlighted that over half of the patients had moderately reduced platelet levels, while a significant portion maintained relatively higher counts (Figure 1).

The analysis of EWS risk categories revealed that the majority of patients (76%) fell into the moderate-risk group, with 90% surviving and 10% succumbing to the illness. In the high-risk group, which comprised 18% of patients, 56% survived while 44% died. The lowrisk group accounted for 6% of patients, all of whom survived. These findings highlighted that a higher EWS was associated with increased mortality, whereas a lower score was linked to more favourable outcomes. Patients who died had significantly higher scores (19.67±4.163) compared to those who survived (11.79±3.216) (P <0.0022) (Table 2).

The analysis of the etiology of febrile thrombocytopenia and EWS risk categories revealed that the majority of patients (60%) were diagnosed with dengue, with 42% falling into the moderate-risk category and 12% into the high-risk group. Other etiologies included rickettsia (8%), sepsis (4%), and other viral infections (8%), each with smaller representations and with most patients categorized to moderate-risk group. The rickettsia group included 6% of patients in the moderate-risk category and 2% in the high-risk category. Each of the sepsis and other viral infection groups had one patient classified as high-risk.

The analysis of platelet count and EWS risk categories showed that the platelet counts in majority of patients (54%) ranged between 50,000 and 100,000/mm³, with 40% in the moderate-risk category and 8% in the highrisk group. Among patients with platelet counts over 100,000/mm³, 20% were classified as moderate risk. Patients with critically low platelet counts (<20,000/ mm³) accounted for 8% of the total, with 6% in the highrisk category. The group with platelet counts between 20,000 and 50,000/mm³ represents 18% of patients, with 14% in the moderate-risk group. Overall, most patients fell into the moderate-risk category across all platelet groups. The P-value for the correlation between platelet group and clinical score risk category was approximately 0.03, indicating a statistically significant association between platelet count and risk category at the 0.05 significance level (Table 3).

The analysis of platelet count groups and clinical outcomes revealed that patients with higher platelet counts generally had more favourable outcomes. Among those with platelet counts between 50,000 and 100,000/ mm³, 54% survived, and no deaths were reported in this group. Similarly, 18% of patients with platelet counts between 20,000 and 50,000/mm³ survived, with no deaths. In the group with platelet counts above 100,000/mm³, 18% survived, while 2% experienced death. However, among patients with platelet counts below 20,000/mm³, all the four patients (8.0%) succumbed. Overall, 90% of patients survived, and 10% experienced death, with mortality more common among those with critically low platelet counts. The P-value for the correlation between platelet group and outcome (Survived or Death) was approximately 0.001. This indicates a statistically significant association between platelet count and patient outcomes at the 0.05 significance level (Table 4).

The analysis of complications across different platelet groups shows that hepatic complications are the most common, especially in patients with platelet counts between 50,000 and 100,000, where 23 patients experienced such issues. Respiratory complications were also significant, particularly in the same platelet group, affecting nine patients. Hematological complications were distributed more evenly, with the highest count (five patients) in the 50,000-100,000 platelet group. Neurological and renal complications were less frequent across all groups, with the majority of these complications occurring in patients with platelet counts below 20,000. Overall, the 50,000-100,000 platelet group had the highest incidence of complications across the board, while the <20,000 group had fewer complications but higher proportions of neurological and renal issues (Figure 2).

Another inference drawn from the study was the association between total number of patients who received platelet transfusions and the organ system complications. Respiratory complications affected 16 patients, with 3 (18.75%) receiving platelets. Renal complications affected 6 patients, with 2 (33.33%) receiving platelets. Hepatic complications were the most common, affecting 46 patients, of whom 4 (8.7%) received platelets. Hematological complications affected 13 patients, with 3 (23.08%) receiving platelets. Finally, neurological complications were rare, affecting only 1 patient, who received platelets (100%). These results indicate that hepatic complications are the most prevalent. The P-value from the Chi-square test was approximately 0.000093. This suggests a statistically significant association between the complications and platelet transfusion status at the 0.05 significance level.

Discussion

The present study aimed to evaluate the utility of an Early Warning Score (EWS) for risk stratification among patients presenting with febrile thrombocytopenia. The results revealed that the EWS, which incorporates platelet counts, vital signs and major organ system complications, was highly effective in identifying patients at high risk of adverse outcomes, including mortality and the need for intensive care unit (ICU) admission. The mean EWS among survivors was significantly lower than non-survivors, underscoring the tool's predictive capacity. These findings align with similar studies, such as Kshirsagar et al., where higher scores were associated with increased mortality and the need for aggressive management strategies.¹² 

In this study, 6% of patients fell into the low-risk category, 76% in the moderate-risk category, and 18% in the high-risk category. The survival rate was 90%, with mortality being significantly higher among patients with elevated EWS. Comparatively, study by Manoj et al., found similar trends, where the EWS accurately predicted the severity of illness and the need for platelet transfusions or ICU care. Both studies emphasized the importance of not relying solely on platelet counts for clinical decisions, which was a consistent finding in our research as well.¹⁹

The majority of our patients had moderate thrombocytopenia with platelet counts between 50,000 and 100,000/µL.20 However, platelet count alone was not a reliable indicator of clinical outcome. Patients with moderate or high platelet counts still exhibited significant complications, reinforcing the findings of Kshirsagar et al., which demonstrated that platelet count should not be the sole determinant for therapeutic interventions such as transfusions. Instead, vital signs and organ system involvement are crucial for making clinical decisions.¹²

Interestingly, the present study identified hepatic complications as the most prevalent among patients with febrile thrombocytopenia, occurring in 92% of cases. This aligns with findings from Swamy et al., where liver function derangements were commonly observed in patients with viral infections such as dengue and leptospirosis. These findings highlight the need for regular monitoring of liver function in patients with febrile thrombocytopenia, particularly in endemic areas.^21,22

Respiratory complications were observed in 32% of the study population, a finding consistent with studies by Sandrock et al., where respiratory distress was a common complication in patients with febrile illnesses. The presence of respiratory and neurological complications was associated with higher Early Warning Score and worse outcomes in our study, reinforcing the need for early recognition and intervention for such complications.^23,24

In contrast to previous studies, our research showed a relatively low incidence of central nervous system (CNS) complications (2%). This may be due to early diagnosis and management, or a smaller sample size compared to larger, multicentre trials. However, studies by Saini et al., have shown that neurological involvement, though less frequent, can significantly worsen patient outcomes when present. Therefore, vigilance for neurological symptoms remains critical in febrile thrombocytopenia.³

This study further demonstrated that unnecessary platelet transfusions can be avoided by using the EWS. Among our patients, 16% received platelet transfusions, that is, 8 out of the 50 patients were transfused, but one (12.5%) transfusion in the low category of EWS could have been avoided if the EWS had been used prospectively. Kshirsagar et al., also similarly reported a decrease in unnecessary transfusions when EWS was used, which helps preserve blood bank resources and reduces patient exposure to transfusion-related complications.¹²

The recommendation from this study is to implement the EWS system in clinical practice to enhance patient outcomes. Patients with moderate and high-risk scores should be monitored more intensively, and decisions regarding platelet transfusions and ICU admission should be made based on the overall clinical presentation, rather than platelet count alone.

Given the findings of this and other similar studies, the recommendation to clinicians is to integrate scoring systems like EWS into routine management protocols for febrile thrombocytopenia. Not only does this improve the accuracy of risk stratification, but it also helps avoid the pitfalls of over-reliance on platelet counts.

Since the study was conducted at a single tertiary care centre with a small sample size of 50 patients, this limits its generalizability to other populations and regions where the prevalence and causes of febrile thrombocytopenia might vary. Moreover, the exclusion of patients with chronic conditions like liver disease or hematological malignancies restricts the applicability of the EWS to more complex cases. Future studies with larger, multicentre populations and longer follow-up durations are recommended to validate the EWS further and enhance its applicability in diverse clinical settings.

Conclusion

Our study highlights the critical importance of utilizing the Early Warning Score (EWS) as a comprehensive risk stratification tool for febrile thrombocytopenia, demonstrating its ability to improve clinical decisionmaking and patient outcomes. By integrating vital signs, platelet counts, and organ system complications, the EWS provides a more reliable method for assessing the severity of illness and guiding interventions such as ICU admissions and platelet transfusions. The findings underscore that, a multifaceted approach is necessary to optimize care, particularly in resource-limited settings. The study contributes to the growing body of evidence supporting the use of scoring systems for febrile thrombocytopenia and emphasizes the need for refined clinical guidelines, and in this case, EWS is most definitely a front runner.

Conflicts of interest

Nil