Study of serum ferritin as diagnostic and prognostic biomarker for severity of sepsis: a hospital-based study

INTRODUCTION

Sepsis is one of the most common causes of death in infants and children in developing countries like India. Sepsis is defined as clinical syndrome which is caused by a dysregulated host response to severe infection.¹ Developing countries with large populations of young children bear the major burden of pediatric sepsis.² Severe sepsis is responsible for >8% of all pediatric intensive care unit (PICU) admissions and causes >4.5 million childhood deaths worldwide per year.³ Sepsis is an advanced stage of infection that overwhelms the host causing capillary leak, hypotension, and/or respiratory failure. Although inflammation is an essential response, the progression to dysregulation of the normal host response leads to widespread tissue injury, immune and microcirculatory dysregulation, and characteristics of systemic inflammatory response syndrome (SIRS). If not treated in a timely manner with appropriate antibiotics and organ supportive treatment, death may occur either due to refractory shock, or due to multiple organ dysfunction syndrome (MODS).

The recognition of sepsis in children is quite challenging and is related to the high prevalence of common seasonal febrile infections, poor specificity of discriminating features, and some capacity of children’s physiology to compensate until shock progresses to advance stage. Ferritin is an iron storage protein highly conserved and widely distributed throughout body.⁴ In critical illness due to sepsis, systemic inflammatory response triggers high levels of proinflammatory cytokines. Proinflammatory cytokines such as interleukin 6, interleukin 8, and tumor necrosis factor stimulate ferritin synthesis.^5,6 Serum ferritin, a positive acute phase reactant increases after an acute inflammatory insult such as sepsis. This study was done as an attempt to understand the usefulness of serial serum ferritin monitoring in predicting the severity of sepsis in pediatric patients.

METHODS

This study was conducted in the Department of Pediatric Medicine, NIMS Medical College and Hospital, Jaipur from October 2022 to September 2023. A total of 80 cases presented with diagnosis of sepsis, severe sepsis, septic shock and sepsis with MODS were enrolled of age 6 months to 18 years.

In the present study, definitions of pediatric sepsis, severe sepsis, and septic shock based on the 2005 International Pediatric Sepsis Consensus Conference7 were followed. These include:

Sepsis

SIRS in the presence of or as a result of suspected or proven infection was sepsis.

Severe sepsis

Sepsis plus one of the following: cardiovascular organ dysfunction or acute respiratory distress syndrome or two or more other organ dysfunctions.

Septic shock

Sepsis and cardiovascular organ dysfunction was septic shock.

Sepsis with MODS

At least two organ failure as a result of severe, persistent and generalized inflammation was sepsis with MODS.

Children whose guardians refused to give positive consent and those who died before taking necessary investigations were excluded. Children affected with hemoglobinopathies, autoimmune disease, hemophagocytic lymphohistiocytosis (HLH), history of earlier repeated blood transfusion and those with chronic systemic diseases such as chronic kidney disease, chronic liver disease were also excluded. This study was a descriptive type of prospective observational study. The clinical and demographic information were recorded on a pre-structured proforma, together with the detail history, physical and detailed systemic examination.

All children were clinically evaluated, investigated, and managed as per PICU protocol, Department of Pediatrics, NIMS Medical College and Hospital. Clinical details, investigation, and progress of sepsis were monitored periodically for each study participant and filled in predesigned case performa. All subjects were monitored clinically on daily basis and more frequently as per requirement of individual case until final outcome. Study subjects were divided into grades of sepsis (sepsis, severe sepsis, septic shock, MODS). The stages of severe sepsis and septic shock were clubbed together into one stage to avoid bias in categorizing them as small margins of discrimination allowed in their differentiation criteria. Progress of sepsis, severe sepsis, septic shock, and sepsis with MODS was followed with serum ferritin levels on day 1, day 3, day 7 or more than 7 days and one week post recovery. Serum ferritin was evaluated and comparison was made in terms of the serum ferritin values in each subgroup and comparing the serum ferritin levels with other groups.

Serum ferritin measurement

Serum ferritin levels were measured by turbidimetric immunoassay using Autoanalyser CST180. The lower and upper limits of detection were 5.2 ng/ml and linearity up to 600 ng/ml, respectively. Possible measures were taken to avoid inter observational and inter instrumental errors.

A written, informed consent was obtained from parents. Clearance from the Departmental Ethics Committee was taken prior to the start of the study. All participants had the option to withdraw from the study anytime during their hospital stay.

All filled questionnaires were checked and coded on Microsoft Office Excel worksheet and any missing data or information was actively searched from patient’s files. Descriptive cross-tabulations were performed to examine for associations. Then a comparison between subgroups was done using Mann-Whiney U test, with p value <0.05 as the cut-off for statistical significance in such comparisons.

RESULTS

In this study, 80 children with diagnosis of sepsis were included. The mean age of the study participants was observed to be 7.75±5.1 years. Four age groups were made to see the common age group of presentation. In this study, 31 (38.8%) children were of age group 6 month-5 years. 22 (27.5%) children related to age group 6-10 years; 24 (30%) children were of 11-15 years age while 3 (3.8%) children belonged to age group of 16-18 years. Maximum of the study participants were in the age group of up to 5 years (38.8%). Among the children, majority of the study participants were males (57.5%), while 42.5% of them were females. Ratio of male to female patients was 1.35:1. In this study, subjects were categorized into three categories on the basis of severity of sepsis. These categories include sepsis, severe sepsis/ septic shock and sepsis with MODS. Sepsis was identified in 37.5% of the total sample. Severe sepsis/septic shock was diagnosed in 50.0% of the total. Sepsis with MODS was observed in 12.5% of the total cases.

*Independent sample Kruskal Wallis test was applied.

The majority of children (50.0%) presented within 3 days of disease onset. A significant portion of the patients (41.3%) presented within 4-7 days. A smaller group of participants (6.3%) had an illness duration of 8- 14 days, and a very limited number (2.5%) had an illness lasting for >14 days before admission. We identified pneumonia as the most common source of infection, representing 35% of the total cases. Gastrointestinal infections were noted in 20% participants. Meningitis was the source of infection for 10.0% of the total. Urinary tract infections were reported in 15.0% of the total cases. An additional category labeled as: others accounted for infections in 20.0% participants. This category included patients of scrub typhus, enteric fever, peritonitis, myocarditis. Blood culture of all participants were sent at the time of admission to look for the causative organism which came positive in nearly 32.5% of the cases. E. coli was the most commonly grown organism in blood culture followed by Klebsiella and Acinetobacter.

In this study, we analyzed the mean serum ferritin level in all categories on days 1, 3, and 7 of admission and 1 week post recovery. On day 1, the mean serum ferritin level was 201.2 ng/ml, with a standard deviation of 90.5 ng/ml, and a median value of 196.0 ng/ml. Day 3 saw an increase in mean serum ferritin levels to 231.0 ng/ml, with a larger standard deviation of 125.8 ng/ml, and a median of 202.0 ng/ml. By day 7, the mean serum ferritin levels decreased to 188.2 ng/ml, with a higher standard deviation of 138.0 ng/ml, and a median value of 153.0 ng/ml. At 1 week after recovery, a further decrease in mean serum ferritin levels to 57.5 ng/ml, with a standard deviation of 40.3 ng/ml, and a median of 52.0 ng/ml was noticed. We compared the levels of serum ferritin on day 1-, 3-, 7- and 1-week post recovery. The median serum ferritin levels increased as sepsis severity worsened, with the highest levels observed in the sepsis with MODS group, followed by the severe sepsis/septic shock group, and the lowest levels in the sepsis group. This pattern suggested that serum ferritin levels may be a relevant marker for assessing the severity of sepsis, with higher levels potentially indicating more severe cases. The p value being less than 0.001 from the Kruskal-Wallis test implies that the observed differences in serum ferritin levels among the three sepsis classifications are unlikely to have occurred by chance. We compared the levels of serum CRP on the day of admission. We observed a significant association between sepsis severity and CRP levels at the time of admission. The highest levels are observed in the sepsis with MODS group, followed by the severe sepsis/septic shock group, and the lowest levels in the sepsis group. The p value being less than 0.001 from the Kruskal-Wallis test indicates that the differences in CRP levels among the three sepsis classifications are statistically significant.

We compared severity of sepsis with the primary site of infection. We found that there were variations in the proportions of sepsis grades depending on the source of infection. The p value of 0.113 suggested that this difference was not statistically significant. As we know, CRP and serum ferritin both are markers of sepsis. In our study, we compared the levels of CRP with serum ferritin at the time of admission. The model showed moderate level of explanatory power (R²=0.5391) and a strong positive correlation (Pearson correlation=0.734) between CRP levels and serum ferritin levels, with statistically significant regression coefficients (p<0.001).

DISCUSSION

This prospective observational study was done with a sample size of eighty subjects, primarily to assess the usefulness of serum ferritin level as diagnostic and prognostic marker of sepsis in pediatric patients. We sought to evaluate if serum ferritin levels correlate with levels of CRP at the time of admission in PICU. We further sought to examine the common source of infection responsible for sepsis in this population. We also wanted to know if severity of sepsis depends on the primary source of infection.

The baseline characteristics in our study showed as median age of participants to be 7.75±5.1 years, with most common age group of 6 month to 5 years and M:F ratio of 1.35:1. Garcia et al was done similar study, in which also most common age range was 2 month to 8 years and male patient were more than female (M:F 2:1).⁸ Study done by Tonial et al also showed similar age group being most common but nearly 1:1 ratio in male and females.⁹ Study done by Rudd et al estimated that 41.5% (20.3 million) of incident sepsis cases and 26.4% (2.9 million) deaths related to sepsis worldwide were among children younger than five years.¹⁰ Younger children had immature immune system, malnutrition and unhealthy dietary habits which make than more vulnerable for infection and sepsis. Our study showed significant male predominance in study subjects. This may be due to ignorance about the health checkups of female children. So that they reach to the health facilities less than the boys. However, no definite causal relationship was found for this male preponderance.

We looked for the duration of fever before seeking health care facility. In our study, most participants had short duration of fever as most patients had fever of <3 days, these patients came to a health care facility in early period of illness as most children presented within 4 days of onset of fever. A study was done by Sarkar et al mean duration of fever more than our study (6.3 days) and this duration was even more in non-survivor patients (6.8 days).¹¹ This showed that as the duration of illness increases, severity of illness may also increase and can lead to fatal outcome. Early presentation to health care facility shows positive attitude of general population to seek treatment.

We looked for the primary site of infection in our patients. Respiratory system was found to be most affected primary site (35%) while gastrointestinal system was second most common affected (20%). In study done by Tonial et al they also found respiratory system to be most common affected syatem and cardiovascular system was second most common.⁹ Respiratory system involvement manifested in form of pneumonia was most common affected system in studies done by Sarkar et al and Kulkarni et al also.^11,12 Previously diarrhoea was the most common cause of childhood infection and mortality while pneumonia was second most common. Over the period of time, significant improvement has been occurred in management of childhood diarrhoea. Now respiratory tract infections are most common cause of mortality in developing countries.

Blood culture is an essential tool for the diagnosis of childhood sepsis. Causative organism can be identified, and antibiotic coverage can be narrowed down after getting a positive blood culture. Blood culture positivity rate in our study was 32.5%. Most common organism grown on blood culture were Gram -ve bacteria in which E. coli was the most common organism. In study by Garcia et al blood culture positivity rate was 33%.⁸ Sarkar et al did comparison of blood culture positivity in survivors and non survivors.¹¹ Survivor group had blood culture positivity rate of 38.2% while non survivors group had positivity rate of 43.3%. In both groups Gram - ve bacteria were more prevalent. In a study done by Lee et al gram-negative organisms, including E. coli, Klebsiella species, and Pseudomonas aeruginosa were frequently identified, especially in neonates, patients with neutropenia, and hospital-acquired sepsis.¹³

In our study, our primary objective was to determine if serum ferritin levels had relation with severity of sepsis. Increased ferritin production is part of the acute phase reaction to infection and induces an iron-deficient milieu by reducing serum available iron.⁵ This process has been hypothesized to have evolved as a defense mechanism to reduce iron available for invading organisms. Serum ferritin levels were compared between different categories at day 1, 3, 7 of admission and 1 week post recovery. In all samplings, we found that serum ferritin levels increased as sepsis severity worsened. In our study, five study participants were expired during study period. All these participants had very high levels of serum ferritin (>600 ng/ml). Garcia et al did study in patients with severe sepsis/septic shock and found that mortality increases with an increase in serum ferritin level.⁸ In their study serum ferritin >500 ng/ml was associated with worse outcomes. Tonial et al did similar study in pediatric patients and found that serum ferritin level >300 ng/ml was associated with 3 fold increase in poor outcome.⁹ Gulrej et al also did similar study in 70 children diagnosed as a case of severe sepsis.¹⁴ In their study, a cut-off value of serum ferritin of 558 ng/ml had a sensitivity of 74.1% and specificity of 67.7% to predict the development of poor outcomes (non-survival). In our study, we could not do ROC curve analysis because there were three categories of sepsis patients. Deaths (6.25% of study subjects) which occurred were included in sepsis with MODS group and the sample size of this group was very small. Proulx et al also analyzed the incidence and outcome of the systemic inflammatory response syndrome (SIRS), sepsis, severe sepsis, and septic shock in a single institution.¹⁵ A total of 1058 admissions were analyzed over a 1-year period. The overall mortality rate for this patient population was 6%, with the majority of deaths occurring in patients with multiple organ dysfunction syndrome (MODS).

In our study, we compared the levels of CRP among different study groups at the time of admission. Values of CRP were highest in MODS group while lowest in sepsis group. The median value of CRP for sepsis group was 26 mg/dl, for severe sepsis/septic shock group 63 mg/dl while for sepsis with MODS group 148 mg/dl. This concluded that level of CRP increases with severity of sepsis. Sarkar et al also compared serum levels of CRP among survivors and non survivors of septic shock patients.¹¹ The values of CRP were not statistically significant among two groups, although CRP level was more in non survivors as compared to survivors.

It is important to acknowledge the limitations of our study. Firstly, due to the small sample size the study was not powered to further analyze the cut off value of serum ferritin to predict mortality. Secondly, we did not analyze various other parameters that may have affected serum ferritin levels such as free iron level, transferrin, lactoferritin. Thirdly, we used CRP and serum ferritin levels as marker of sepsis. These both are indirect markers of infection because these increase in inflammation which occurs in sepsis. These both can be falsely elevated in inflammatory conditions other than sepsis. Finally, our institute is the only referral center and the major hospital that provides primary to tertiary care for children in this region. The geographic location of our hospital may have led to an enrollment bias.

CONCLUSION

Serum ferritin levels can be used as a marker of sepsis both in terms of severity of sepsis and mortality predictor. Serum ferritin values can be followed to look for the response to treatment as improving patients will show significant decrease in ferritinemia. More specific and low-cost tests need to be developed to look for severity of sepsis in children.

Funding: No funding sources

Conflict of interest: None declared

Ethical approval: The study was approved by the Institutional Ethics Committee

REFERENCES

1. 1. Goldstein B, Giroir B, Randolph A; International Consensus Conference on Pediatric Sepsis. International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatric Crit Care Med. 2005;6(1):2-8.
2. 2. Global report on the epidemiology and burden of sepsis: current evidence, identifying gaps and future directions. Geneva: World Health Organization; 2020. Available at: https://www.who.int/publications/i/item/9789240010789.
3. 3. Davis AL, Carcillo JA, Aneja RK, Deymann AJ, Lin JC, Nguyen TC, et al. American college of critical care medicine clinical practice parameters for hemodynamic support of pediatric and neonatal septic shock. Crit Care Med. 2017;45(6):1061-93.
4. 4. Jurado RL. Iron, infections, and anemia of inflammation. Clin Infect Dis. 1997;25:888-95.
5. 5. Diaz J, Arribas JM, Vallina E, Maradona JA, Hevia C, Blanco F. Acute- phase reactants in sepsi. Rev Clin Esp. 1992;191:473-7.
6. 6. Darveau M, Denault AY, Blais N, Notebaert E. Bench-to bedside review: iron metabolism in critically ill patients. Crit Care. 2004;8:356-62.
7. 7. Singer M. The Third International consensus definitions for sepsis and septic shock (sepsis-3). JAMA. 2016;315(8):801-10.
8. 8. Garcia PC, Longhi F, Branco RG, Piva JP, Lacks D, Tasker RC. Ferritin levels in children with severe sepsis and septic shock. Acta Pediatric. 2007;96(12):1829-31.
9. 9. Cristian T, Caroline AD, Andrades G, Crestani F, Paulo R, Bruno F, et al. Prediction of poor outcomes for septic children according to ferritin levels in a middle-income setting. Pediatr Crit Care Med. 2020;21(5):259-66.
10. 10. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, TsoiD, Kievlan DR, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet. 2020;395(10219):200-11.
11. 11. Sarkar M, Roychowdhury S, Zaman MA, Raut S, Bhakta S, Nandy M. Can serum ferritin be employed as prognostic marker of pediatric septic shock and severe sepsis. J Pediatr Critic Care. 2021;8(1):20.
12. 12. Kulkarni MS, Shukla DA. Study of role of serum ferritin in predicting outcome in children with severe sepsis at a tertiary hospital. Med Pulse Int J Pediatr. 2021;19(3):48-52.
13. 13. Lee CY, Chen PY, Huang FL. Microbiologic spectrum and susceptibility pattern of clinical isolates from the pediatric intensive care unit in a single medical center-6 years' experience. J Microbiol Immunol Infect. 2009;42(2):160-5.
14. 14. Shaikh GN, Ramamoorthy JG, Parameswaran N. Serum ferritin for predicting outcome in children with severe sepsis in the pediatric intensive care unit. Indian Pediatr. 2022;59:939-42.
15. 15. Proulx F, Fayon M, Farrell CA. Epidemiology of sepsis and multiple organ dysfunction syndrome in children. Chest. 1996;109:1033-7.