Trans R Soc Trop Med Hyg 2021; 0: 1–9

https://doi.org/10.1093/trstmh/trab135 Advance Access publication 0 2021

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

The immature platelet fraction, a predictive tool for early recovery
from dengue-related thrombocytopenia: a prospective study

Original Article
Visula Abeysuriyaa,∗ , Suranjith L. Seneviratnea,b , Primesh de Mela , Choong Shi Hui Claricec , Chandima de Mela ,
Lal Chandrasenaa , Christina Yipd , Eng-Soo Yapd , and Sanjay de Melc

a
Nawaloka Hospital Research and Education Foundation, Nawaloka Hospitals PLC, Colombo 00200, Sri Lanka; b Institute of Immunity
and Transplantation, Royal Free Hospital and University College London, London NW3 2QG, UK; c Department of Haematology -
Oncology, National University Cancer Institute, National University Health System Singapore, 5 Lower Kent Ridge Road 119074,
Singapore; d Department of laboratory Medicine, National University Health System, 5 Lower Kent Ridge Road 119074, Singapore
∗ Corresponding author: Tel: [+94] 77 326 555 2; E-mail: [email protected]

Received 8 December 2020; revised 16 May 2021; editorial decision 16 August 2021; accepted 23 August 2021

Background: There is a paucity of predictive factors for early recovery from thrombocytopenia related to dengue.
The immature platelet fraction (IPF%) is reflective of megakaryopoiesis and may correlate with recovery from
dengue-related thrombocytopenia. Our objective was to assess the predictive value of IPF% on days 2 and 3 of
illness for recovery from dengue-related thrombocytopenia.
Methods: A prospective study was conducted among patients with dengue admitted to our institution
(Nawaloka Hospital PLC) from December 2019 to October 2020. Dengue was diagnosed based on positive non-
structural antigen 1 or IgM. IPF% data were extracted from the Sysmex-XN-1000 automated hematology ana-
lyzer. Clinical data were obtained from electronic medical records. Statistical analyses were performed using
SPSS version 20.
Results: We included 240 patients. An IPF% on day 2 of illness of >7.15% had a sensitivity of 80.0% and speci-
ficity of 70.4% for prediction of platelet recovery (defined as platelet count ≥60×109 /L) on day 7 of illness.
An IPF% of >7.25% on day 3 of illness had a sensitivity of 88.9% and specificity of 47.1% for predicting platelet
recovery >60×109 /L on day 8 of illness. The IPF% was significantly lower in patients with severe dengue. Platelet
recovery was observed within 48 h after the peak IPF% was reached, regardless of severity.
Conclusion: We propose that IPF% values on days 2 and 3 of illness are a promising predictive tool for early
recovery from dengue-related thrombocytopenia.

Keywords: dengue, immature platelet fraction, megakaryopoiesis, predictive tool, recovery, thrombocytopenia

Introduction timing and severity of dengue-related thrombocytopenia is an

active field of research.6–8
Dengue is the most prevalent arboviral infection worldwide, with Biomarkers predictive of early platelet recovery are equally
estimates of >390 million infections per year.1,2 Mortality due to important as they may help to identify patients who could be
dengue typically occurs in the context of dengue hemorrhagic managed in the outpatient setting. This is an important consider-
fever and dengue shock syndrome, which occur in 2–5% of ation given the limited healthcare resources in many of the coun-
patients.3 Thrombocytopenia is a common feature of severe tries where dengue is endemic.2 Immature platelets (also known
dengue (SD) and may contribute to hemorrhagic complica- as reticulated platelets) represent the platelets most recently
tions.4 While thrombocytopenia in dengue is believed to be released into the circulation from the bone marrow.9,10 Their
predominantly immune mediated, complications of SD such as larger size and higher cytoplasmic RNA content enables auto-
disseminated intravascular coagulation may also contribute.4 mated hematology analyzers (including the Sysmex XN and XE
No specific treatment has shown conclusive clinical benefit series as well as Abbot Cell-Dyn Sapphire and Mindray BC-6800)
in dengue-related thrombocytopenia, hence management is to distinguish them from mature platelets.9 Immature platelets
mainly supportive.4,5 Identification of biomarkers to predict the

© The Author(s) 2021. Published by Oxford University Press on behalf of Royal Society of Tropical Medicine and Hygiene.
All rights reserved. For permissions, please e-mail: [email protected]

1
V. Abeysuriya et al.

Table 1. Demographic and clinical characteristics of the study Table 1. Continued.

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

population
Variable Number (%)
Variable Number (%)
Pleural effusion 14 (5.8)
Age category (y) Respiratory depression 14 (5.8)
<30 80 (33.3) Altered conscious level 3 (1.2)
31–40 59 (24.6)
1 Comorbidities were diagnosed previously by a consultant physi-
41–50 47 (19.6)
51–60 36 (15.0) cian based on clinical history, examination and laboratory
>61 18 (7.5) investigations. Cut-off values for laboratory investigations for diag-
Gender nosis of comorbidities was based on the WHO STEPS survey guide-
lines (WHO STEPS surveillance manual, 26 January 2017). Some
Male 111 (46.2)
patients had more than one comorbidity.
Female 129 (53.8) 2 NS1, dengue virus non-structural protein 1; IgM, dengue virus-
Comorbidities1 specific immunoglobulin M; IgG, dengue immunoglobulin G. Details
None 188 (78.3) of assay protocols in supplementary data.
Diabetes mellitus 23 (9.6) 3 WHO and Special Program for Research and Training in Tropi-

Hypertension 17 (7.1) cal Diseases. Handbook for clinical management of dengue; 2012.
Hyperlipidemia 12 (5.0) (Refer to Supplementary Table 1 for details of classification.)
* RHC, right hypochondriac region.
Other 8 (3.3)
Day of illness on admission
1 27 (11.3)
2 86 (35.8)
3 103 (42.9) as a proportion of total platelets (immature platelet fraction
4 20 (8.3) [IPF%]) and the absolute immature platelet count are generated
5 4 (1.7) as research parameters at no extra cost by these analyzers.9,11
Dengue serologic test positivity2 The normal range for the IPF% has been reported at approxi-
NS1 240 (100) mately 1–7% in healthy adults.11,12 IPF% has been shown to be
NS1 and IgM 49 (20.4) a useful marker of thrombopoietic activity and has utility in the
NS1, IgM and IgG 11 (4.6) differentiation of peripheral destruction from marrow failure as a
NS1 and IgG 13 (5.4) cause of thrombocytopenia.13
Dengue severity (WHO 2009 classification)3 Achieving a rising IPF% has recently been associated with
(refer to Supplementary Table 1 for details of classification) platelet count recovery in dengue.14–16 These studies were, how-
Dengue without warning signs 197 (82.1) ever, conducted on relatively small sample sizes and did not
Dengue with warning signs 29 (12.1) directly evaluate the predictive power of the IPF% during the early
Severe dengue 14 (5.8) stages of infection for platelet recovery. We therefore sought to
r Dengue shock syndrome 9 (64.28) prospectively assess the utility of IPF% on admission as a predic-
r Severe bleeding 4 (28.57) tive factor for recovery of thrombocytopenia related to dengue in
r Severe organ dysfunction 8 (57.14) a larger cohort of patients. We also evaluated differences in IPF%
Common symptoms and signs based on clinical severity and probable primary vs probable sec-
Symptoms ondary dengue.
Fever 240 (100)
Nausea 61 (25.4)
Arthralgia/myalgia 216 (90.0)
Headache 195 (81.2)
Materials and Methods
Retro-orbital pain 203 (84.6) Study design
Abdominal pain 39 (16.2)
A prospective study was conducted among patients with dengue
Vomiting 42 (17.5)
admitted to Nawaloka Hospital (NH), Colombo, Sri Lanka, from
Diarrhea 13 (5.4)
December 2019 to October 2020.
Shortness of breath 17 (7.1)
Signs
Macular skin rash 37 (15.4) Study participants and data collection
Petechial rash 37 (15.4)
We screened 303 patients with clinical features suspicious for
Bleeding manifestations 41 (17.1)
dengue according to the Sri Lankan national guidelines.17 The
RHC* tenderness 41 (17.1)
study protocol for screening and selection of suspected cases
Epigastric tenderness 38 (15.8)
are summarized in Supplementary Data 3A and 3B. We included
Ascites 42 (17.5)
240 patients aged 18–80 y with dengue. The diagnosis of
Hepatomegaly 34 (14.2)
dengue was confirmed by positivity for the dengue non-structural

2
Transactions of the Royal Society of Tropical Medicine and Hygiene

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

Figure 1. Receiver operator characteristic (ROC) curve analysis of the immature platelet fraction (IPF%) on day 2 of illness as a predictive factor for a
platelet count of >60 × 109 /L on day 7 of illness.

Figure 2. Receiver operator characteristic (ROC) curve analysis of the immature platelet fraction (IPF%) on day 3 of illness as a predictive factor for a
platelet count of >60 × 109 /L on day 8 of illness.

antigen 1 (NS1) and/or dengue IgM. All patients were also Patients with a known history of hematologic malignancy,
tested for dengue IgG. NS1 was assessed using a qualitative immunosuppression, HIV positive status or known infection with
lateral flow immunochromatographic assay (FD standard diag- other viral pathogens were excluded. Clinical parameters, includ-
nostic, Korea). Dengue IgG and IgM were tested using dengue ing presenting symptoms, vital signs, results of other blood inves-
IgG/IgM Combo Rapid Test kits (CTK biotech test kit, USA). tigations and clinical progress (specifically the development of
Blood samples were collected on admission by venipuncture SD), were recorded by a study team member during the patient’s
and tested for NS1, IgM and IgG, along with full blood count admission. Severity of dengue was classified according to WHO
(FBC), renal and liver function tests. Following admission, patients guidelines (Supplementary Table 1).18 Sociodemographic data
with dengue without warning signs had FBC (which includes were obtained from patients’ medical records. There is no stan-
IPF), liver and renal function tests performed daily. Patients dard platelet threshold to define ‘recovery’ from thrombocytope-
with dengue with warning signs and SD had FBC performed nia related to dengue. For the purposes of our study we used a
twice a day. The individual reports were authenticated by two platelet count of 60×109 /L to define recovery, given that the risk
technologists. of spontaneous bleeding would be very low at this threshold.19

3
V. Abeysuriya et al.

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

Figure 3. Variation of the immature platelet fraction (IPF%) and platelet count (109 /L) over time in patients stratified according to severity of dengue
(LOESS curve).

Immature platelet fraction data (216/240 [90.0%]). All patients were NS1 positive while those
positive for NS1 and IgM comprised 20.4% (49/240). NS1, IgM
IPF% data were extracted from the Sysmex XN 1000 automated
and IgG positive patients accounted for 4.6% (11/240) of cases.
FBC analyzer (Sysmex Kobe Japan). Sysmex XN-1000 analyzers
Fever was the most common symptom followed by arthral-
employ a carefully designed gating system in the optical (fluo-
gia/myalgia and retro-orbital pain. Patients with SD and dengue
rescence) reticulocyte/platelet channel to reliably quantitate the
with warning signs comprised 5.8% and 12.1% of all cases,
IPF.20 Details of the IPF assay and related quality control pro-
respectively, and no deaths occurred during the study period.
tocols are explained in detail in the supplementary data (assay
Patients were managed according to WHO 2009 guidelines18
protocol 2).
and none required platelet transfusion. The clinical and demo-
graphic characteristics of the study population are summarized in
Data analysis Table 1.
We evaluated whether IPF% on days 2 and 3 of dengue is
Data were entered into EXCEL worksheets, checked manually and predictive of early recovery from dengue-related thrombocytope-
logically, then corrected where necessary. Descriptive statistics nia. We performed ROC curve analysis of the IPF% readings on
were derived and expressed as measures of central tendency days 2 and 3 of illness to determine if they can predict a platelet
and frequencies. ANOVA was used to compare averages. Cut-off count of >60×109 /L on day 7 or 8 of illness respectively (Figures 1
values were derived from receiver operator characteristic (ROC) and 2). Fourteen patients had a platelet count of >60×109 /L
curve analysis. Data were analyzed using Statistical Package for before day 7; their results were not included in the prediction
Social Sciences 20 (SPSS 20.0, Chicago, IL, USA) and STATA version analysis. For financial reasons, these patients were transferred
12 (TX, USA). p<0.05 was considered significant. to another institution for further follow-up on recovery of their
platelet counts to >60×109 /L. Hence we did not have access to
their platelet counts following recovery. We found that an IPF%
on day 2 of illness >7.15% had a sensitivity of 80.0% and speci-
Results ficity of 70.4% for prediction of platelet recovery on day 7 of ill-
We enrolled 240 patients (male; n=111 [46.2%]) with a median ness (AUC: 89.4 [95% CI 76.7 to 95.1]). Furthermore, on day 3 of
age of 35 y. The median day of fever on admission was 3. The illness, an IPF% of >7.25% had a sensitivity of 88.9% and speci-
majority of patients had no comorbidities (188/240 [78.3%]) and ficity of 47.1% for predicting platelet recovery on day 8 of illness
were admitted at, or earlier than 3 d from the onset of fever (AUC: 78.3 [95% CI 66.9 to 89.7]). IPF% readings on subsequent

4
Transactions of the Royal Society of Tropical Medicine and Hygiene

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

Figure 4. (A) Immature platelet fraction (IPF%) and platelet count (109 /L) according to severity of dengue among patients admitted on day 2 of illness
(n=86). (B) IPF% and platelet count (109 /L) according to severity of dengue infection among patients admitted on day 3 of illness (n=103).

days such as days 4 and 5 of illness had lower predictive power We demonstrated variation of IPF% and platelet count
than the readings on days 2 and 3 of illness. (109 /L) according to severity of dengue among patients
Next, we examined the association between IPF% and platelet admitted on day 2 (n=86) and day 3 (n=103) of illness
count during the first 7 d of illness. The mean peak IPF% read- (Figure 4A,B). The IPF% started to drop on days 6 and 7 of
ings were reached on day 5 of illness in patients with dengue illness once platelet recovery was established. The mean IPF%
without warning signs (DWWS) and dengue with warning signs readings, as well as platelet counts on all 6 days were signifi-
(DWS). The mean peak IPF% was reached on day 6 of illness in cantly lower in patients with SD and DWS compared with DWWS
patients with SD. Platelet count recovery occurred 48 h after the (p<0.0001) (Figure 4A,B). A similar pattern of IPF% and platelet
IPF% reached its peak, regardless of disease severity. The rela- count variation over the course of illness was seen in patients
tionship between IPF% and platelet count for each category of admitted on days 1 or 4 of illness (Figure 5A,B and Supplementary
dengue is summarized in Figure 3. Table 4A,B).

5
V. Abeysuriya et al.

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

Figure 5. (A) Dengue with or without warning signs. Variation of the immature platelet fraction (IPF%) during days 3, 4, 5 and 6 of illness. Optical
(fluorescence) platelet scattergrams with forward scattered light on the y-axis and fluorescence on the x-axis. (B) Severe dengue. Variation of the
IPF% during days 3, 4 and 5 of illness. Optical (fluorescence) platelet scattergrams with forward scattered light on the y-axis and fluorescence on the
x-axis.

6
Transactions of the Royal Society of Tropical Medicine and Hygiene

Table 2. (A) Immature platelet fraction (IPF%) and platelet count (109 /L) according to the serological subtype of dengue infection among

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

patients admitted on day 2 of illness (n=86)

Day 2 Day 3 Day 4 Day 5 Day 6 Day 7

Variable Mean±SD (n) Mean±SD (n) Mean±SD (n) Mean±SD (n) Mean±SD (n) Mean±SD (n)
Immature platelet fraction (IPF%)
NS1 6.19a ±1.51 (86) 9.60a ±2.48 (86) 13.44a ±2.91 (86) 11.41a ±2.05 (84) 9.33a ±1.71 (81) 7.92a ±1.18 (79)
NS1 and IgM 4.46b ±1.30 (19) 6.21b ±1.31 (19) 9.82b ±1.51 (19) 9.64b ±1.31 (18) 8.18a ±1.05 (16) 7.04a ±1.11 (14)
NS1, IgM and IgG 3.21c ±1.10 (4) 4.18c ±1.03 (4) 6.18c ±2.03 (4) 7.51c ±1.04 (4) 7.27b ±1.10 (4) 6.41b ±1.22 (3)
NS1 and IgG 4.11b ±1.43 (4) 5.31b ±1.73 (4) 7.70d ±2.93 (4) 7.90c ±1.77 (4) 6.45c ±1.02 (4) 6.02b ±0.91 (4)
p-value1 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 =0.001
F(3,109) =13.16 F(3,109) =20.24 F(3,109) =21.03 F(3,106) =11.95 F(3,101) =7.52 F(3,96) =6.43

Day 2 median (n) Day 3 median (n) Day 4 median (n) Day 5 median (n) Day 6 median (n) Day 7 median (n)
IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th)
Platelet count (109 /L)
NS1 113 (86) 102 (86) 95 (86) 84 (84) 87 (81) 98 (79)
(101–122) (93–113) (78–113) (62–101) (68–101) (79–114)
NS1 and IgM 96 (19) 75 (19) 66 (19) 56 (18) 66 (16) 78 (14)
(84–106) (65–86) (61–78) (44–69) (42–83) (64–86)
NS1, IgM and IgG 84 (4) 70 (4) 63 (4) 44 (4) 58 (4) 75 (3)
(76–88) (67–76) (59–71) (40–49) (47–66)
NS1 and IgG 81 (4) 61 (4) 52 (4) 45 (4) 46 (4) 64 (4)
(79–88) (58–65) (45–58) (39–51) (40–51) (57–66)
p-value2 <0.001 <0.0001 <0.0001 <0.0001 <0.0001 <0.007

1 ANOVA; post hoc test: Tukey. Normal distribution was assessed by Kolmogorov–Smirnov test p>0.05.
2 KruskalWallis test.
Superscript a,b,c and d demonstrated the significant difference of each mean value of IPF within each day.

We subsequently evaluated differences in IPF% between prob- >60 × 109 /L on day 7 and day 8 of dengue illness. This IPF%
able primary and secondary dengue. Probable primary dengue reading is significantly higher than that of patients with non-
was defined as NS1 and NS1+IgM positive cases, while prob- specific viral fever at a similar time point in their symptoms. We
able secondary dengue comprised NS1, IgM and IgG posi- also demonstrate that IPF% differs significantly based on clini-
tive cases or NS1 and IgG positive cases. We demonstrated cal severity and probable primary vs secondary dengue. Other
variation of IPF% and platelet count (109 /L) according to the platelet indices, including mean platelet volume, platelet crit
serological subtype of dengue among patients admitted on and platelet distribution width, have been evaluated as poten-
day 2 (n=86) and day 3 (n=103) of illness (Table 2A,B). Patients tial biomarkers of dengue severity.21,22 None of these, however,
with probable primary dengue had a significantly higher mean showed correlation with recovery of thrombocytopenia or clinical
IPF% on all days of illness compared with those with secondary manifestations.
dengue (p<0.001). Correspondingly, patients with probable pri- Given the utility of IPF% in other disorders associated with
mary dengue (in particular, the group positive for NS1 only) thrombocytopenia (such as immune thrombocytopenic purpura
showed higher platelet counts compared with those with proba- and aplastic anemia),13 it has also become a field of active study
ble secondary dengue on all 6 days (p<0.0001). A similar pattern in dengue. Serial IPF% readings during the course of dengue have
of variation in IPF% and platelet count according to serological shown some correlation (albeit not statistically significant) with
subtype of dengue was seen in patients admitted on days 1 and platelet counts in a pediatric cohort.15 An increase in IPF% >10%
4 of illness (Supplementary Table 5A,B). was shown to be associated with platelet recovery within 72 h in
Demographic and clinical characteristics of dengue NS1 and studies from India and Thailand.14,23–25 Two studies have demon-
IgM negative patients are presented in Supplementary Data 3C. strated that the IPF% on admission had a positive correlation
The working diagnosis for these patients was that of a non- with platelet counts at 24 and 48 h from admission.26,27 Recov-
specific viral fever. They all recovered uneventfully and were dis- ery from dengue-related thrombocytopenia typically occurs at
charged within 3–4 d of presentation. The hematologic charac- approximately day 9 of illness.4 The majority of our patients were
teristics of this cohort are presented in Supplementary Data 3D. admitted on days 2 and 3 of illness, hence day 6 of admission
The mean IPF% for these patients was 1.98±0.23 on day 2 and correlates with the timing of platelet recovery for the bulk of our
1.91±0.34 on day 3 of illness. cohort.
We present the largest study evaluating IPF% as a predictive
factor for platelet recovery in dengue. In keeping with previous
Discussion studies, all our patients showed platelet recovery within 48 h of
the IPF% reaching its peak value. The analysis of serial IPF% over
We have shown that an IPF%>7.15% on day 2 of illness and 6 days in our cohort provides a more robust assessment than pre-
>7.25% on day 3 of illness are predictive of a platelet count of vious studies, where two or three time points were assessed. The

7
V. Abeysuriya et al.

Table 2. (B) Immature platelet fraction (IPF%) and platelet count (109 /L) according to the serological subtype of dengue infection among

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

patients admitted on day 3 of illness (n=103)

Variable Day 3 Mean±SD (n) Day 4 Mean±SD (n) Day 5 Mean±SD (n) Day 6 Mean±SD (n) Day 7 Mean±SD (n) Day 8 Mean±SD (n)
Immature platelet fraction (IPF%)
NS1 6.49a ±1.96 (103) 9.05a ±2.62 (103) 12.42a ±3.65 (100) 10.77a ±2.47 (95) 8.99a ±1.79 (90) 7.68a ±1.34 (84)
NS1 and IgM 5.66b ±1.77 (29) 7.92b ±2.31 (29) 10.22b ±3.51 (29) 9.57b ±1.71 (25) 8.38a ±1.15 (21) 7.37a ±1.14 (20)
NS1, IgM and IgG 4.21c ±1.17 (7) 5.78c ±1.53 (7) 7.58c ±2.13 (7) 7.54c ±1.84 (7) 6.67b ±1.46 (6) 6.27b ±1.20 (5)
NS1 and IgG 3.31b ±1.34 (9) 5.14b ±1.23 (9) 6.80d ±1.93 (9) 7.15c ±1.77 (8) 7.05c ±1.01 (8) 6.22b ±0.81 (6)
p-value1 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 =0.008
F(3,144) =11.04 F(3,144) =10.62 F(3,141) =12.10 F(3,131) =10.48 F(3,121) =6.85 F(3,111) =4.41

Day 3 median (n) Day 4 median (n) Day 5 median (n) Day 6 median (n) Day 7 median (n) Day 8 median (n)
IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th) IOQ (25th to 75th)
Platelet count (109 /L)
NS1 100 (103) 98 (103) 87 (100) 77 (95) 70 (90) 88 (84)
(91–113) (82–115) (64–102) (54–89) (50–89) (56–99)
NS1 and IgM 86 (29) 64 (29) 65 (29) 57 (25) 60 (21) 75 (20)
(77–96) (53–73) (51–89) (38–69) (41–77) (54–98)
NS1, IgM and IgG 60 (7) 45 (7) 37.05 (7) 34 (7) 38 (6) 57 (5)
(46–71) (39–57) (21–54) (26–39) (24–46) (41–67)
NS1 and IgG 87 (9) 64 (9) 66.70 (9) 57 (8) 59 (8) 84 (6)
(79–98) (56–75) (51–77) (43–61) (49–64) (63–98)
p-value2 <0.001 <0.001 =0.001 <0.0001 =0.03 =0.025

1 ANOVA; post hoc test: Tukey. Normal distribution was assessed by Kolmogorov–Smirnov test p>0.05.
2 KruskalWallis test.
Superscript a,b,c and d demonstrated the significant difference of each mean value of IPF within each day.

predictive power of the IPF% on days 2 and 3 of illness for platelet and day 8 of illness. Given that our study is from a single cen-
recovery on day 7 or day 8 of illness was not assessed in any of ter in Sri Lanka, these findings need to be validated in indepen-
the prior studies. We also believe that the ROC analysis in our dent cohorts outside south Asia. Further studies are also required
study provides a more rigorous assessment of predictive power to evaluate differences in IPF% among dengue serotypes and in
compared with the correlation coefficient that was used in other a larger number of patients with SD. Given that IPF% data can
studies. also be generated by other hematology analyzers,9 further stud-
Our finding of significantly lower IPF% in patients with SD and ies are required to validate our findings using these platforms.
DWS suggests that marrow suppression is more pronounced in Finally, the correlation between IPF% and coagulation parame-
these patients compared with those with mild disease.4 The find- ters, as well as fibrinolysis in SD, should be evaluated in future
ing of lower IPF% in patients with probable secondary dengue studies.
supports the hypothesis that immune mechanisms are respon-
sible for the suppression of megakaryopoiesis in these patients.4
The correlation of IPF% with platelet function would be of signif-
icant interest and should be evaluated in future studies. Supplementary data
Our findings suggest that patients with an IPF% >7 on Supplementary data are available at Transactions online.
days 2 and 3 of illness may be considered for close outpatient
monitoring rather than admission. The average length of hospi-
tal stay for patients with dengue ranges from 3 to 7 d.28 The aver-
age cost of hospitalization ranged from US$216–609 for pediatric
Authors’ contributions: ESY, CY, SDM, SLS and VA conceptualized the
cases to US$196–866 for adults in Sri Lanka.29 Biomarkers of early study; VA and PDM collected the data; VA, SDM, SLS and CY analyzed the
platelet recovery such as IPF% could therefore result in significant data; VA, SDM, CSHC, ESY and CY wrote the manuscript; VA, LC, CDM, SLS
cost reductions through identification of patients who could avoid and SDM conducted a critical review and editing of the manuscript. All
admission or be discharged early. The limitations of our study the authors read and approved the final version of the manuscript. VA
include the fact that we do not have dengue serotype data for and SDM are guarantors of the paper.
our patients, as well as the relatively small number of patients
with SD. Given the relatively young age of our cohort, these find- Acknowledgements: We acknowledge the assistance given by the Direc-
ings need to be replicated in older patients with dengue. tor/General Manager and the management of the Nawaloka Hospital PLC,
Colombo, Sri Lanka. We also thank the staff of the Medical Records Office
and Computer Division unit of Nawaloka Hospital, Colombo. We would
Conclusions
also like to thank Sysmex Asia Pacific regional branch for their kind assis-
We propose that the IPF% values on days 2 and 3 of illness are tance. Finally, we thank all patients and next-of-kin for providing the nec-
a reliable predictor of platelet recovery to >60 × 109 /L on day 7 essary information required for the study.

8
Transactions of the Royal Society of Tropical Medicine and Hygiene

Funding source: The hematology laboratory at NH received XN reagents 14 Dadu T, Sehgal K, Joshi M, et al. Evaluation of the immature platelet
and funds from Sysmex Asia Pacific Pte Ltd for conduct of the study. fraction as an indicator of platelet recovery in dengue patients. Int J

Downloaded from https://academic.oup.com/trstmh/advance-article/doi/10.1093/trstmh/trab135/6367993 by Copyright Clearance Center user on 23 September 2021

This study was funded by a research grant from the Sri Lanka Medical Lab Hematol. 2014;36(5):499–504.
and Dental Association of the UK. 15 Amrutha BS, Adarsh E, SreeKrishna Y, et al. Immature platelet frac-
tion in children infected with dengue fever. Int J Contemp Pediatr.
Competing interests: None declared. 2018;6:5.
16 Ong-Misa Maria MS, Garcia Robert DJ, Uy-Aragon MJ, et al.
Ethical approval: Ethical approval for this study was obtained from the
Relationship between immature platelet fraction and platelet
ethical committee of Nawaloka Hospital, Colombo, Sri Lanka.
count among pediatric patients with dengue fever: a prospec-
Data availability: The data that support the findings of this study are tive cross-sectional study. 10th World Pediatric Congress,
available on request from the corresponding author. 2017.
17 Ministry of Health - Sri Lanka. Guidelines on Management of Dengue
Fever &. Dengue Haemorrhagic Fever in adults. Revised. Ministry of
Health - Sri Lanka; 2012.
References 18 WHO. Dengue guidelines for diagnosis, treatment, prevention and
1 Simmons CP, Farrar JJ, van Vinh Chau N, et al. Dengue. N Engl J Med. control : new edition. Geneva, Switzerland: World Health Organization;
2012;366(15):1423–32. 2009.
2 Bhatt S, Gething PW, Brady OJ, et al. The global distribution and burden 19 Slichter SJ. Relationship between platelet count and bleeding risk
of dengue. Nature. 2013;496(7446):504–7. in thrombocytopenic patients. Transfus Med Rev. 2004;18(3):153–
3 Guzman MG, Gubler DJ, Izquierdo A, et al. Dengue infection. Nat Rev 67.
Dis Prim. 2016;2(1):16055. 20 Sysmex XN Series XN-1000 Instruction for Use. English | Electromag-
4 De Azeredo EL, Monteiro RQ, De-Oliveira Pinto LM. Thrombocytope- netic Interference | Electromagnetic Compatibility. 2014.
nia in dengue: Interrelationship between virus and the imbalance 21 Sharma K, Yadav A. Association of mean platelet volume with severity,
between coagulation and fibrinolysis and inflammatory mediators. serology & treatment outcome in dengue fever: prognostic utility. J
Mediators of Inflammation, 2015;2015:313842. https://doi.org/10. Clin Diagnostic Res. 2015;9:EC01–3.
1155/2015/313842.
22 Khatri S, Sabeena S, Arunkumar G, Mathew M. Utility of platelet param-
5 Schexneider KI, Reedy EA. Thrombocytopenia in dengue fever. Curr eters in serologically proven dengue cases with thrombocytopenia.
Hematol Rep. 2005;4:145–8. Indian J Hematol Blood Transfus. 2018;34(4):703–6.
6 Abeysuriya V, Choong CSH, Thilakawardana BU, et al. The atypical lym- 23 Kumar VV, Senthilkumaran S, Thirumalaikolundusubramanain P.
phocyte count: a novel predictive factor for severe thrombocytopenia Immature platelet fraction in dengue cases. Int J Infect Dis.
related to dengue. Trans R Soc Trop Med Hyg. 2020;114(6):424–32. 2016;45:443.
7 Clarice CSH, Abeysuriya V, de Mel S, et al. Atypical lymphocyte 24 Lühn K, Simmons CP, Moran E, et al. Increased frequencies of
count correlates with the severity of dengue infection. PLoS One. CD4+CD25 high regulatory T cells in acute dengue infection. J Exp
2019;14(5):e0215061. Med. 2007;204(5):979–85.
8 de Mel S, Thilakawardana BU, de Mel P, et al. Triple positivity for non- 25 Chuansumrit A, Apiwattanakul N, Sirachainan N, et al. The use
structural antigen 1, immunoglobulin M and immunoglobulin G is pre- of immature platelet fraction to predict time to platelet recov-
dictive of severe Thrombocytopenia related to dengue infection. J Clin ery in patients with dengue infection. Paediatr Int Child Health.
Virol. 2020;129: 104509. 2020;40(2):124–8.
9 Buttarello M, Mezzapelle G, Freguglia F, Plebani M. Reticulated platelets 26 Renuka BG, Upadhya S. Immature platelet fraction (IPF) as an indica-
and immature platelet fraction: clinical applications and method lim- tor of platelet recovery in dengue fever QR Code. Int J Biomed Res.
itations. Int J Lab Hematol. 2020;4:363–70. 2019;10:5305.
10 Ault KA, Rinder HM, Mitchell J, et al. The significance of platelets with 27 Ahmad W, Lubna Z, Mohsin A, et al. Study of platelet indices in dengue
increased RNA content (reticulated platelets): a measure of the rate fever with thrombocytopenia and correlation of immature platelet
of thrombopoiesis. Am J Clin Pathol. 1992;98(6):637–46. fraction (IPF) with platelet recovery. Arch Hematol Case Reports Rev.
11 Bat T, Leitman SF, Calvo KR, et al. Measurement of the absolute imma- 2020;5(1):1–5.
ture platelet number reflects marrow production and is not impacted 28 Mishra S, Chopra D, Jauhari N, et al. A study on length of stay
by platelet transfusion. Transfusion. 2013;53(6):1201–4. and its predictors among dengue patients in a tertiary care insti-
12 Pons I, Monteagudo M, Lucchetti G, et al. Correlation between imma- tute in Lucknow. Int J Community Med Public Health. 2019;6(11):
ture platelet fraction and reticulated platelets. Usefulness in the etiol- 4870.
ogy diagnosis of thrombocytopenia. Eur J Haematol. 2010;85:158–63. 29 Thalagala N, Tissera H, Palihawadana P, et al. Costs of dengue con-
13 Jeon K, Kim M, Lee J, et al. Immature platelet fraction: a useful marker trol activities and hospitalizations in the public health sector dur-
for identifying the cause of thrombocytopenia and predicting platelet ing an epidemic year in urban Sri Lanka. PLoS Negl Trop Dis. 2016;
recovery. Med. 2020;99(7):e19096. 10(2):e0004466.