Clinical efficacy of immunoglobulin on the treatment of severe fever with thrombocytopenia syndrome: a retrospective cohort study

Summary

Background

Optimal treatment strategy for severe fever with thrombocytopenia syndrome (SFTS) remained unknown. We aimed to evaluate the efficacy of intravenous immunoglobulin (IVIG) on SFTS.

Methods

A retrospective cohort study was conducted based on medical records of the laboratory-confirmed SFTS patients hospitalized during 2010–2020 in the 154th hospital, China. A 1:1 propensity score matching with age, sex, the interval from symptom onset to admission, presence of chronic viral hepatitis, diabetes and disease severity was performed between Non-IVIG group (supportive therapy) and IVIG group (IVIG plus supportive therapy). The matching variables were adjusted to compare the case fatality rates (CFRs), viral load and laboratory parameters between the two groups. Risk ratio (RR) and 95% confidence interval (CI) were reported.

Findings

Totally 2219 SFTS patients were recruited. CFRs were significantly higher in 1051 patients in IVIG group than 1168 patients in Non-IVIG group (19.0% vs. 4.6%, RR = 4.30, 95% CI 3.12–5.93). The difference remained significant after matching (17.2% vs. 5.1%, RR = 4.02, 95% CI 2.71–5.97). The CFR of IVIG group was significantly higher in all age groups, two IVIG therapy delay groups and two therapy duration groups compared to that of Non-IVIG group (all P < 0.05). IVIG therapy was related to higher viral loads and reduced counts of lymphocytes, T cells, CD4⁺ T cells and natural killer cells in the blood (all P < 0.05).

Interpretation

No obvious efficacy of IVIG in saving life or improving outcome of SFTS was observed. Caution is needed for clinical physicians to continue prescribing IVIG for SFTS patients.

Funding

Natural Science Foundation of China.

Research in context.

Evidence before this study

Severe fever with thrombocytopenia syndrome (SFTS) remained one of the high mortality viral hemorrhagic fevers, with optimal treatment strategy unknown. We aimed to evaluate the efficacy of intravenous immunoglobulin (IVIG) on SFTS.

We performed a search on PubMed for articles published using the search terms “immunoglobulin” or “IVIG” and “severe fever with thrombocytopenia syndrome” or “SFTS” or “bunyavirus” or “SFTSV”, with no language or time restrictions. Altogether 30 papers were retrieved, among which only one was a research article that studied the effect of IVIG in treating SFTS patients by retrospective analysis, but the sample was small. In the other papers, one was a review including the clinical treatment to cope with the disease, three case reports reported the clinical efficacy of IVIG in treating clinical SFTS patients, and the rest were unrelated to this study topic. Few systematic analyses have investigated the effects of therapy by using case-control study design or cohort design, especially in a relatively large sample. As the drug has been already used widespread in the endemic regions, there are urgent need to clarify the clinical efficacy of IVIG treatment on SFTS patients.

Added value of this study

Based on a retrospective cohort study on laboratory confirmed SFTS patients, this work had demonstrated adverse effect from IVIG treatment, manifested by a higher mortality rate, delayed recovery of laboratory abnormalities, together with higher viremia and suppressed immune response than those receiving no IVIG. Neither advantage was observed when patients were further stratified and analyzed by age, gender, timing of therapy initiation, therapy duration or therapy dose. Additional immunological and metabonomic analysis was further performed to explore the underlying mechanism of the adverse response post IVIG administration. The adverse effects of IVIG have been previously demonstrated as suppressing the host immune response via multiple mechanisms. Most of these potential adverse effects had been identified in the current study, manifested by significantly reduced counts of T cell, B cell and NK cell in SFTS patients after IVIG therapy, on the premise of comparable demography and baseline clinical features. In that case, the advantages of IVIG might be overwhelmed by the adverse effects, since the dysfunctional immune cells were among the key potential strategy of SFTSV to compromise host immune.

The metabonomic analysis also revealed metabolic changes that might be related to the adverse effects of IVIG usage. It’s notable that metabolism of tryptophan was significantly altered and of central in the metabolic pathways in patients after receiving IVIG therapy.

Implications of all the available evidence

The findings of this study provide persuasive data that increase the degree of certainty that IVIG offer no obvious efficacy in saving life or improving outcome of SFTS. Caution is needed for clinical physicians to continue prescribing IVIG, particularly when other efficacious treatments are available for SFTS.

Introduction

Severe fever with thrombocytopenia syndrome (SFTS) is an emerging tick-borne zoonosis caused by a novel Bandavirus of the Phenuiviridae family, Dabie bandavirus, also known as SFTS virus (SFTSV).¹ Since its emergence nearly a decade ago, SFTS has affected a rapidly growing number of people, with a consistently expanding geographic distribution, mostly in China, Korea and Japan.^2,3 SFTS is characterized by acute onset of fever accompanied by thrombocytopenia and leukopenia.⁴ In most cases, SFTSV infection presents as a mild disease but can develop into life-threatening illness, which occurs at 7–10 days after the onset of symptoms in over 10% of the patients.⁵ Despite all these public health significance and clinical significance, efficient therapeutics against SFTSV are unavailable.

Clinical trials of antivirals (favipiravir), case control study of repurposed drugs (calcium channel blocking), case reports on the monoclonal antibodies targeting SFTSV had been reported to be effective in reducing fatality.6, 7, 8 Considering cytokine storm as one of the key points of SFTS pathogenic process, clinicians also applied steroids and plasma exchange to suppress the inflammatory response of the patients.^9,10 Other supportive adjuvant therapy, for example, intravenous immunoglobulin (IVIG), and recombinant human granulocyte colony-stimulating factor were also administered,¹¹ as recommended by the expert consensus on diagnosis and treatment suggestions for SFTS.¹²

IVIG is a blood preparation isolated from healthy plasma donors consisting of over 95% of IgG and trace amounts of IgA or IgM.¹³ It has been originally used as substitutional therapy for primary or acquired immunodeficiencies, e.g., in treating idiopathic thrombocytopenic purpura, the Guillain-Barré syndrome, chronic inflammatory demyelinating polyneuropathy, myasthenia gravis, corticosteroid-resistant dermatomyositis, and Kawasaki's syndrome etcetera in order to improve plasma IgG concentrations and provide passive immunity.¹⁴ The effects are exerted via modulating the expression and function of Fc receptors, interference with the activation of complement and the cytokine network, provision of antiidiotypic antibodies, activation and differentiation of T cells and B cells, etc.¹⁴

Based on its complex function, particularly the efficacy in reducing inflammation, IVIG has long been administered in treating viral infections, including Ebola virus, human immunodeficiency virus, influenza virus and West Nile virus (WNV).15, 16, 17, 18 Different research including clinical trials, animal or in vitro experiments and observational studies yielded inconsistent results on the effect of IVIG treatment in these diseases, with beneficial effect reported in improving prognosis in studies treating influenza B,¹⁷ while others showing ambiguous effect of IVIG, for example, on treating Crimean Congo hemorrhagic fever and WNV infection.^18,19 Even without clear consensus in its efficacy, when emerging infectious disease was firstly discovered to infect human being, drugs with unproven efficacy might be prescribed in clinical practice. The most recent example is the currently ongoing COVID-19 pandemic, although the routine usage of IVIGs was not recommended, it was empirically applied, yielding inconsistent results from several clinical studies, systematic review and meta-analysis.^20,21

Such a similar manner was seen in the treatment of SFTS. The expert consensus in China recommended IVIG usage as adjuvant therapy, however provided no explicit recommendations regarding dose or indicators of its usage.¹² Although several studies reported advantage in saving life, they are only from one single observational study with small sample size and several case reports.^9,22,23 There was large discrepancy of the patient characteristics and regimens among studies, thus the initial evidence of the clinical efficacy might not be replicated and was debatable. As the drug was already in widespread use in the endemic regions, there are urgent need to clarify the clinical efficacy of IVIG treatment on SFTS patients. Here by using an observational, retrospective cohort study, we were designed to evaluate the efficacy of IVIG in hospitalized patients with SFTSV infection compared to standard of care (SOC) patients. The immunological and metabonomic alterations post IVIG administration were further explored.

Methods

Study design and participants

A retrospective cohort study was conducted in the 154th hospital, a designated hospital for SFTS therapy in Xinyang city, Henan province, China, which was the most severe SFTS endemic area in China. All the studied patients were hospitalized during 2010–2020 and confirmed with SFTSV infection according to guideline released by National Health Commission of China.²⁴ A laboratory confirmation of SFTSV infection was made by the criteria as previously described, i.e., positive detection of SFTSV RNA from reverse transcription-polymerase chain reaction (RT-PCR) or real-time RT-PCR in the blood of patients.²⁵

A medical record review was performed by a group of trained physicians to collect the information from the medical database using a standardized format, which included demography, symptoms and signs, laboratory test results and treatment regimens during the entire hospitalization. The data were entered into an EpiData database and further reviewed for accuracy and consistency by a group of epidemiologists. Patients were excluded if (1) aged <18 years; (2) received favipiravir; (3) with hospitalization duration less than three days or failed to be sampled. The comorbidities considered for the present analysis included hypertension (ICD-10 I10.X02), diabetes mellitus (both type I and type II, ICD-10 E14.8), chronic virus hepatitis (both HBV and HCV, ICD-10 B18.951), chronic heart diseases (CHD, ICD-10 I51.900), cerebral ischemic stroke (ICD-10 I64.X04), chronic obstructive pulmonary diseases (ICD-10 J44.900), pulmonary tuberculosis (ICD-10 B90.901), and cancer (ICD-10 C00–C97).

Laboratory tests

Blood samples were collected from laboratory confirmed patients on admission into hospital and during hospitalization. SFTSV loads in the serum was quantified by real-time RT-PCR targeting the same gene segments as used for the diagnosis, which were expressed as copies/mL. Seven hematologic indicators were tested using Sysmex XE 2100 (Kobe, Japan). The count of total lymphocyte and its subsets were determined in the peripheral blood using a MultiTEST IMK kit (BD Biosciences) following the manufacturer’s instructions.²⁶ Metabolites level in the serum were determined by LC-MS/MS analyses. The list of tested indicators and the testing methods were summarized in the Supplementary Methods and Supplementary Table S1.

Treatment

Since the efficacy of IVIG on treating SFTS was uncertain, and due to its potential side effects, it was administrated to a subset of patients according to the individual suggestion of attending clinicians and the willingness of the patients and their family, after they were fully informed about the advantage and risk of the treatment. Dose of 250 mg/(Kg·day) was prescribed for at least five days or until discharge from hospital or SFTSV RNA was undetectable, except that the patients or their family requested termination of IVIG therapy. In addition to IVIG, antiviral treatment (ribavirin), symptomatic therapy, and supportive treatment (including hepatoprotective treatment, plasma/platelet transfusion, immunomodulatory therapy with Xuebijing or recombinant human granulocyte colony-stimulating factor), and antibiotics or glucocorticoids were prescribed if deemed necessary. The patients who received any of the above-mentioned treatments were defined as the Non-IVIG group and those who also received IVIG therapy were defined as the IVIG group.

Outcomes

The primary outcome was survival or death, which information was retrieved from medical records or obtained by following up the patients who discontinued therapy or had been discharged from hospital because of adverse clinical progression or for economic reasons. The follow-up was performed within one month of discharge by phone call to determine their final outcome (death or survival). Secondary outcomes included consecutively evaluated viral loads and laboratory parameters, including white blood cell (WBC) count, platelet (PLT) count, aspartate transaminase (AST), alanine transaminase (ALT), creatine kinase (CK), lactate dehydrogenase (LDH), and albumin (ALB). The subgroup CFR were further calculated stratified by age (≤60 years, 60–70 years and >70 years), sex, the interval from symptom onset to admission (≤5 days and >5 days) and bleeding or neurological symptoms (presence and absence).

Ethics statement

The study was approved by the ethics committee of 154th Hospital (154YYLL-2015-10). Informed consent was obtained before sampling and data collection.

Statistical analysis

Continuous variables were expressed as median and interquartile range (IQR), for which the comparison between the IVIG and the Non-IVIG group was performed using nonparametric Mann–Whitney U test before matching. Wilcox rank-sum test after matching was used for matched subjects. Categorical variables were described as counts and percentages, for which the inter-group comparison was made using the χ² test or Fisher’s exact test.

To control for confounding, a propensity score model (PSM) in a 1:1 ratio was used to match and compare patients treated and not treated with IVIG during the first 48 h of treatment. Variables included in the models were chosen based on their univariate relation to the treatment exposure or mortality in either the study cohort or based on prior literature.⁴ The included variables were age, sex, the interval from symptom onset to admission, underlying CVH, DM, disease severity (defined as the presence of bleeding/neurological symptoms before therapy started). Mortality outcomes between cases treated and not treated with IVIG were compared using propensity-matched logistic regression models to calculate risk ratio (RR) with associated 95% confidence interval (CI) after adjusting age, sex, the interval from symptom onset to hospital admission, and comorbidity. Demographic and clinical data were available for all included patients, but there were some data missing for viral load results. Patients who had given over two tests of viral loads were used for another PSM with matching the above variables and viral load before IVIG therapy.

Subgroup analysis was likewise performed by therapy duration (≤3 days and >3 days), therapy delay (≤6 days and >6 days). RR and 95% CI in subgroups were also estimated by logistic regression models after adjusting the same covariates as above except the stratified variable. Assuming the CFR is 8% in the Non-IVIG group and 16% in the IVIG group, the sample size needs to be no less than 690, 345 in each group, to reach the power of 0.9 and alpha of 0.05, which was calculated by PASS 2021.

The generalized estimating equation (GEE) was constructed to compare the inter-group difference in viral loads and laboratory measurements which were continuously evaluated over time. Generalized linear model was used to examine the difference of immune cell counts and metabolites between the Non-IVIG and the IVIG groups after adjusting the basic characteristics (age, sex, the interval from symptom onset to admission and comorbidity), and a box-cox transformation was conducted when the data was not normal-distributed. The multivariate analysis of serum metabolome for differential metabolites screening were conducted by orthogonal partial least square discriminate analysis (OPLS-DA), using software SIMCA (version 14.1, Sartorius Stedim Biotech, Umea, Sweden), in which significant differential metabolites between the IVIG group and the Non-IVIG group were selected according to the variable importance in the projection (VIP) values, P value and fold change (FC) (VIP>1, P < 0.05, FC > 1 or FC < 0.5). The enrichment analysis was based on MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/) and the differential pathway was searched on the Kyoto Encyclopedia of Genes and Genomes database (KEGG; https://www.genome.jp/kegg/pathway.html) (P < 0.05).

All the statistical analyses were performed using R version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria) and STATA 17 (StataCorp LLC, College Station, TX77845, USA). A two-sided P < 0.05 was considered as statistically significant.

Role of funders

The funders did not play any role in the study design, data collection, statistical analysis, interpretation, writing or submission of the manuscript. The corresponding authors had complete access to the data and hold the final responsibility for the decision to submit for publication.

Results

Patients and baseline analysis

From January 2012, to June 2020, 2940 patients with laboratory-confirmed SFTSV infection were admitted into the 154th hospital. Among them 721 patients were excluded due to having received favipiravir (n = 503), with <3 days of hospitalization duration or inadequate sampling (n = 202), <18 years old (n = 15), and unknown clinical outcome (n = 1); thus, the remaining 2219 patients were included, of whom 1051 (47.4%) received IVIG therapy (Fig. 1 and Table 1). Totally 789 patients treated with IVIG were successfully matched to 789 patients receiving no IVIG. In this matched cohort, all the covariates (age, sex, the interval from symptom onset to admission, presence of comorbidities, and clinical symptoms on hospital admission) were well balanced except for two of the subjective symptoms (feeble and chills) (Table 1).

Fig. 1.

Flowchart of the study design. SFTS, Severe fever with thrombocytopenia syndrome; PSM, propensity score matching; IVIG, intravenous immunoglobulin; CVH, chronic viral hepatitis; DM, diabetes mellitus.

Table 1.

Baseline characteristics of SFTS patients in the study.

Characteristics	Before PSM				After PSM
	Total (N = 2219)	Non-IVIG (n = 1168)	IVIG (n = 1051)	P	Total (N = 1578)	Non-IVIG (n = 789)	IVIG (n = 789)	P
Age, median (IQR)	62 (53–70)	61 (50–69)	64 (57–70)	<0.001	63 (54–70)	64 (54–70)	62 (53–69)	0.061
Age groups, n (%)				<0.001				0.133
≤60 years	949 (42.8)	574 (49.1)	375 (35.7)		648 (41.1)	307 (38.9)	341 (43.2)
60–70 years	779 (35.1)	361 (30.9)	418 (39.8)		572 (36.2)	289 (36.6)	283 (35.9)
>70 years	491 (22.1)	233 (19.9)	258 (24.5)		358 (22.7)	193 (24.5)	165 (20.9)
Sex, male, n (%)	889 (40.1)	460 (39.4)	429 (40.8)	0.491	675 (42.8)	320 (40.6)	355 (45.0)	0.075
The interval from symptom onset to admission, median (IQR)	5 (4–7)	5 (4–7)	5 (4–7)	0.571	5 (4–7)	5 (4–6)	5 (4–7)	0.629
Comorbidity, n (%)	746 (33.6)	364 (31.2)	382 (36.3)	0.010	534 (33.8)	266 (33.7)	268 (34.0)	0.915
Hypertension	248 (11.2)	134 (11.5)	114 (10.8)	0.640	175 (11.1)	99 (12.5)	76 (9.6)	0.065
COPD	203 (9.1)	99 (8.5)	104 (9.9)	0.247	152 (9.6)	75 (9.5)	77 (9.8)	0.865
DM	132 (5.9)	57 (4.9)	75 (7.1)	0.025	89 (5.6)	39 (4.9)	50 (6.3)	0.230
CVH	218 (9.8)	91 (7.8)	127 (12.1)	0.001	145 (9.2)	66 (8.4)	79 (10.0)	0.257
CHD&CIS	136 (6.1)	68 (5.8)	68 (6.5)	0.525	100 (6.3)	52 (6.6)	48 (6.1)	0.679
TB	21 (0.9)	11 (0.9)	10 (1.0)	0.981	15 (1.0)	8 (1.0)	7 (0.9)	0.795
Cancer	14 (0.6)	6 (0.5)	8 (0.8)	0.462	12 (0.8)	5 (0.6)	7 (0.9)	0.562
General symptoms, n (%)	2216 (99.9)	1166 (99.8)	1050 (99.9)	1.000	1575 (99.8)	787 (99.7)	788 (99.9)	1.000
Fever	2210 (99.6)	1162 (99.5)	1048 (99.7)	0.610	1570 (99.5)	783 (99.2)	787 (99.7)	0.288
Chills	275 (12.4)	163 (14.0)	112 (10.7)	0.019	186 (11.8)	109 (13.8)	77 (9.8)	0.012
Headache	298 (13.4)	156 (13.4)	142 (13.5)	0.915	213 (13.5)	102 (12.9)	111 (14.1)	0.507
Dizziness	461 (20.8)	251 (21.5)	210 (20.0)	0.382	320 (20.3)	161 (20.4)	159 (20.2)	0.900
Feeble	2129 (95.9)	1103 (94.4)	1026 (97.6)	<0.001	1513 (95.9)	743 (94.2)	770 (97.6)	0.001
Myalgia	1806 (81.4)	953 (81.6)	853 (81.2)	0.794	1278 (81.0)	640 (81.1)	638 (80.9)	0.898
Lymphadenectasis	1247 (56.2)	658 (56.3)	589 (56.0)	0.889	897 (56.8)	452 (57.3)	445 (56.4)	0.722
Gastrointestinal symptoms, n (%)	2102 (94.7)	1099 (94.1)	1003 (95.4)	0.158	47 (3.0)	18 (2.3)	29 (3.7)	0.103
Anorexia	1716 (77.3)	902 (77.2)	814 (77.5)	0.900	16 (1.0)	8 (1.0)	8 (1.0)	1.000
Nausea	1333 (60.1)	706 (60.4)	627 (59.7)	0.705	6 (0.4)	2 (0.3)	4 (0.5)	0.683
Vomit	611 (27.5)	341 (29.2)	270 (25.7)	0.065	4 (0.3)	2 (0.3)	2 (0.3)	1.000
Diarrhea	460 (20.7)	233 (19.9)	227 (21.6)	0.338	19 (1.2)	6 (0.8)	13 (1.6)	0.106
Bleeding symptoms, n (%)	74 (3.3)	23 (2.0)	51 (4.9)	<0.001	4 (0.3)	0 (0.0)	4 (0.5)	0.133
Melena	21 (0.9)	10 (0.9)	11 (1.0)	0.644	1485 (94.1)	737 (93.4)	748 (94.8)	0.240
Hemoptysis	6 (0.3)	2 (0.2)	4 (0.4)	0.590	1206 (76.4)	600 (76.0)	606 (76.8)	0.722
Haematemesis	5 (0.2)	2 (0.2)	3 (0.3)	0.906	948 (60.1)	481 (61.0)	467 (59.2)	0.472
Hematuria	35 (1.6)	9 (0.8)	26 (2.5)	0.001	443 (28.1)	232 (29.4)	211 (26.7)	0.239
Petechial	10 (0.5)	0 (0)	10 (1.0)	0.002	330 (20.9)	166 (21.0)	164 (20.8)	0.901
Neurological symptoms, n (%)	164 (7.4)	63 (5.4)	101 (9.6)	<0.001	87 (5.5)	43 (5.4)	44 (5.6)	0.912
Dysphoric	23 (1.0)	8 (0.7)	15 (1.4)	0.085	12 (0.8)	6 (0.8)	6 (0.8)	1.000
Convulsion	69 (3.1)	31 (2.7)	38 (3.6)	0.193	38 (2.4)	20 (2.5)	18 (2.3)	0.743
Blurred Mind	85 (3.8)	30 (2.6)	55 (5.2)	0.001	47 (3.0)	21 (2.7)	26 (3.3)	0.459
Coma/Somnolence	46 (2.1)	14 (1.2)	32 (3.0)	0.002	20 (1.3)	11 (1.4)	9 (1.1)	0.653

There were no mission data.

IQR, inter-quartile range. SFTS, severe fever with thrombocytopenia syndrome; PSM, propensity score matching; IVIG, intravenous immunoglobulin; COPD, chronic obstructive pulmonary diseases; DM, diabetes mellitus; CVH, chronic viral hepatitis; CIS, cerebral ischemic stroke; CHD, chronic heart diseases; TB, tuberculosis.

Therapy regimen

Overall, IVIG was administered in 127 patients with duration ≥5 days and in 662 patients with duration of <5 days, resulting in a median duration of 3 (IQR 2–4) days. Among all recorded SOC, hepatoprotective drugs, ribavirin, plasma transfusion, and Xuebijing were administered with significantly higher frequencies in the IVIG group than in the Non-IVIG group (P < 0.05, Supplementary Table S2).

Effect of IVIG therapy on CFR

Before PSM, the crude CFRs were significantly higher in the IVIG group than in the Non-IVIG group (19.0%, 200/1051 vs. 4.6%, 54/1168, adjusted RR [aRR] = 4.30, 95% CI 3.12–5.93, P < 0.001) (Fig. 2 and Supplementary Table S3). After PSM, this IVIG vs. Non-IVIG difference in CFR remained significant (17.2% vs. 5.1%, aRR = 4.02, 95% CI 2.71–5.97, P < 0.001). When subgroup analysis was performed, the IVIG-related difference remained significant among each age group (10.3% vs. 1.6%, aRR = 6.57, 95% CI 2.53–17.08 for ≤60 years; 20.8% vs. 6.2%, aRR = 3.93, 95% CI 2.25–6.89 for 60–70 years, 25.5% vs. 8.8%, aRR = 3.52, 95% CI 1.91–6.52 for >70 years), sex (18.3% vs. 6.6%, aRR = 3.56, 95% CI 2.09–6.04 for male and 16.4% vs. 4.1%, aRR = 5.47, 95% CI 3.16–9.45 for female), and the interval from symptom onset to admission (16.6% vs. 4.1%, aRR = 5.24, 95% CI 3.09–8.89 for ≤5 days and 18.2% vs. 6.7%, aRR = 3.37, 95% CI 1.95–5.54 for >5 days) with all P < 0.05. However, discrepancy was observed regarding disease severity. For patients without severe disease, those treated with IVIG had a significantly higher CFR than those without IVIG usage (16.3% vs. 4.2%, aRR = 4.83, 95% CI 3.18–7.35), while for patients with severe disease, the difference failed to show significance (27.1% vs. 15.8%, P = 0.072) (Fig. 2, Supplementary Tables S3 and S4).

Fig. 2.

The effect of IVIG therapy on CFR of SFTS patients. Datapoints show odds ratios and error bars show 95% confidence interval. The red color represents P < 0.05 and the black color represents P ≥ 0.05. Risk ratios (RR) were calculated by multivariate conditional logistic regression models among patients after matching, in which comorbidity was adjusted. ∗P < 0.05 and P values were calculated by Pearson chi-square test. SFTS, Severe fever with thrombocytopenia syndrome; PSM, propensity score matching; IVIG, intravenous immunoglobulin; CFR, case fatality rate; SFTS, Severe fever with thrombocytopenia syndrome.

Considering the side effects of ribavirin in treating SFTS, we made a stratified analysis by its usage. Significantly increased CFR was observed for IVIG within both subgroups of ribavirin receiving (aRR = 4.90, 95% CI 1.95–12.30) and non-receiving (aRR = 4.31, 95% CI 2.83–6.55) (Supplementary Table S5).

We made further stratified comparison regarding the therapy regimens of IVIG. Compared with controls with no IVIG therapy, we observed a significantly increased CFR in patients with early IVIG therapy (less than 6 days post disease onset) (14.4%, 56/389 vs. 5.1%, 20/389, P = 0.001), as well as late IVIG therapy (later than 6 days post disease onset) (20.0%, 80/400 vs. 5.0%, 20/400, P < 0.001) (Fig. 3 and Supplementary Table S6). Patients who received IVIG therapy later than 6 days from symptom onset had a significantly higher CFR than those receiving IVIG within 6 days among all patients (20.0% vs. 14.4%, P = 0.022) and those aged 60–70 years (23.7% vs. 18.1%, P = 0.031) (Supplementary Table S7).

Fig. 3.

The effect of IVIG therapy on CFR in SFTS patients after therapy delay and therapy duration stratified. a: Therapy delay ≤6 days; b: Therapy delay >6 days; c: Therapy duration ≤3 days; d: Therapy duration >3 days. Therapy delay refers to the interval from symptom onset to receive IVIG therapy and therapy duration refers to the duration of IVIG therapy. Datapoints show odds ratios and error bars show 95% confidence interval. The red color represents P < 0.05 and the black color represents P ≥ 0.05. Risk ratios (RR) were calculated by multivariate conditional logistic regression models among patients after matching, in which comorbidity was adjusted. ∗P values were calculated by Pearson χ² test. SFTS, Severe fever with thrombocytopenia syndrome; IVIG, intravenous immunoglobulin; CFR, case fatality rate; SFTS, Severe fever with thrombocytopenia syndrome.

For patients who received IVIG therapy either with duration >3 days or ≤3 days, no advantage in decreasing CFR was observed compared with the Non-IVIG therapy group (20.3% vs. 3.6% for the >3 days group; 16.0% vs. 5.6% for the ≤3 days group; both P < 0.05) (Fig. 3 and Supplementary Table S8). Moreover, in patients aged ≤ 60 years, those receiving IVIG therapy with duration >3 days had a significantly higher CFR than those with duration ≤ 3 days (15.5% vs. 8.2%, P = 0.049, Supplementary Table S9).

Secondary outcomes

Totally 114 SFTS patients were included after matching for age, sex, the interval from symptom onset to admission, CVH, DM, disease severity and viral load before IVIG therapy, of which the baseline information was comparable between the IVIG group (n = 57) and the matched Non-IVIG group (n = 57) (Supplementary Table S10). GEE analysis revealed a significantly higher viremia in patients receiving IVIG therapy than those receiving no IVIG across the disease course (P < 0.05, Fig. 4). In the premise of comparable levels of 7 laboratory indicators that were evaluated before the 6th day from symptom onset, we observed significantly increased level in the IVIG group than in the non-IVIG group for AST, ALT, CK, and LDH beginning at the 6th day from symptom onset through GEE analysis, while significantly lower level in the IVIG group than in the non-IVIG for PLT, WBC and ALB during the same period (all P < 0.001, Supplementary Fig. S1). A longer duration for WBC, PLT, CK and ALB to return to normal value was observed in patients who received IVIG therapy compared with the Non-IVIG group (Supplementary Fig. S1).

Fig. 4.

Kinetics of viral loads from the 4th day in SFTS patients with or without IVIG treatment. Datapoints are median values and error bars show inter-quartile range. P values were calculated by generalized estimating equation. The numbers of patients who contributed to the at-risk population at each time point are shown under the x-axis. SFTS, Severe fever with thrombocytopenia syndrome; IVIG, intravenous immunoglobulin.

Immunological cell counts

Overall, 109 patients (median age 62 years, IQR 55–69; 47 males; 14 deceased) were evaluated for the lymphocyte and its subset counts of blood samples that were collected at 7–9 days from symptom onset. The IVIG group (n = 41) and the Non-IVIG group (n = 68) were comparable for age, sex and interval from symptom onset to sample collection (Supplementary Table S11). The counts of lymphocytes, T cells, CD4⁺ T cells, and natural killer (NK) cells were significantly lower in the IVIG group as compared with the Non-IVIG group (all P < 0.05, Fig. 5 and Supplementary Table S12). The counts of B cells and CD8⁺ cells, percentage of CD3-ζ chain expression in T cell and CD4⁺/CD8⁺ T cells ratio were also lower in the IVIG group, although with no statistical significance attained (Supplementary Table S12).

Fig. 5.

Comparison of immune cell counts between the Non-IVIG and IVIG groups at acute stage of SFTS patients. a: Lymphocyte count (10⁹/L); b: T cell count (/μL); c: CD4⁺ T cell count (/μL); d: NK cell count (/μL). The lines are median values and points show values of each observation. ∗P < 0.05 and P values were calculated in general linear model. IVIG, intravenous immunoglobulin; SFTS, severe fever with thrombocytopenia syndrome; NK cell, natural killer cell.

Metabolic profile

Totally 42 patients (21 receiving IVIG and 21 receiving no IVIG with comparable clinical characteristics were subject to the metabolic analysis (Supplementary Table S13). Totally 534 metabolites were evaluated on samples collected 1–4 days after receiving IVIG, with comparable interval from symptom onset to sample collection between the two groups. A clear clustering of patients with or without IVIG therapy was obtained, supported by the OPLS-DA of supervised models (Fig. 6a). Differential metabolites that were classified into super classes included organic acid and its derivatives, fatty acyls, amino acid and its derivatives, nucleotide and its derivates, phospholipid and bile acids, all listed as the top ones (Fig. 6b). There were 5 key differential metabolites were observed increased including L-kynurenine, taurohyocholic acid sodium salt, tauro-alpha-muricholic acid sodium salt, taurochenodeoxycholic acid sodium salt and taurodeoxycholic acid sodium salt (Fig. 6c–g). Based on all the differential metabolites, two key differential metabolic pathways were screened, tryptophan metabolism pathway and glycine, serine and threonine metabolism pathway (KEGG enrichment analysis, P < 0.05) (Fig. 6h). L-kynurenine was the differential metabolite involved in tryptophan metabolism pathway (Fig. 6c).

Fig. 6.

Metabolic profile between the Non-IVIG and IVIG groups in SFTS patients. a: The OPLS-DA; b: The super class of significant differential metabolites; c: The level of L-lynurenine; d: The level of taurohyocholic acid sodium salt; e: The level of taurodeoxycholic acid sodium; f: The level of tauro-alpha-Muricholic acid sodium salt; g: The level of taurochenodeoxycholic acid sodium salt; h: The KEGG pathway mapping of differential metabolites. IVIG, intravenous immunoglobulin; SFTS, severe fever with thrombocytopenia syndrome; OPLS-DA, orthogonal partial least square discriminate analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Discussion

IVIG is a widely used treatment option in severe viral infection, even though evidence indicating the ineffectiveness has been accumulating. In China, Korea and other Asian countries where IVIG has been applied to treat SFTS, the effectiveness had been described in several case reports.¹¹ In Korea, two SFTS cases with neurological complications were reported to recover after a combined treatment with IVIG and corticosteroid.⁹ Another case report recorded one pediatric case with SFTS treated with a combination of steroids, IVIG, and ribavirin, who recovered without sequelae.²³ There was also a retrospective cohort study reporting beneficial effects of IVIG therapy for SFTS patients, however, with a small sample and the conclusion was debatable.²² Until recently, there was no convincing case-control study or cohort study to investigate the effects of IVIG therapy. Our study, based on a retrospective analysis of a laboratory confirmed SFTS cohort, had demonstrated adverse effects from IVIG treatment, manifested by a higher mortality rate, delayed recovery of laboratory abnormalities, together with higher viremia and suppressed immune response than those receiving no IVIG. Neither advantage was observed when patients were further stratified and analyzed by age, gender, timing of therapy initiation, therapy duration or therapy dose.

The mechanism underlying the adverse effects of IVIG might be explicated from two perspectives. Differing from influenza, SFTSV infection was restricted to rural areas in the geographical clusters in China. It’s likely that IVIG applied in the clinical practice contains no specific antibodies against SFTSV, thus failing to offer protection in SFTSV infection. Actually, the SFTSV antibody elicited in the asymptomatic individuals with previous SFTSV infection were measured at low level and without prolonged duration, even in the high-endemic region.²⁷ The similar lack of beneficial effects of IVIG have been observed in the treatment of Crimean Congo hemorrhagic fever and WNV infection, both of which are endemic disease with limited distribution.17, 18, 19

Moreover, the adverse effects of IVIG have been previously demonstrated as suppressing the host immune response via multiple mechanisms. IVIG contain anti-TCR, anti-CD4/CD8, anti-HLA antibodies and soluble CD4/CD8 etc., which bind to antigen-presenting cells and competitively inhibit the immune activation of autoreactive T cells.^28,29 IVIG is also suggested to inhibit the activation and proliferation of T cells. Most of these potential adverse effects had been identified in the current study, manifested by significantly reduced counts of T cell, B cell and NK cell in SFTS patients after IVIG therapy, on the premise of comparable demography and baseline clinical features. In that case, the advantage of IVIG might be overwhelmed by the adverse effect, since the dysfunctional immune cells were among the key potential strategy of SFTSV to compromise host immune.

The metabonomic analysis also revealed metabolic change that might be related to the adverse effects of IVIG usage. It’s notable that metabolism of tryptophan was significantly altered and of central in the metabolic pathways in patients after receiving IVIG therapy. L-kynurenine, derived from the catabolism of tryptophan by the enzymes indoleamine 2,3-dioxygenase 1 and tryptophan 2,3-dioxygenase 2, has been proven for its immunosuppression effect, such as via the pro-apoptotic effect on NK cells.30, 31, 32 This result might suggest the potential correlation between increased L-kynurenine and depressed immune response such as the reduction of T cell and NK cell following IVIG usage. Moreover, the taurocholic acid was reported to be elevated for the accumulation of bile acid in the liver and systemic circulation, causing damage to the bile duct and liver cells through triggering inflammation, necrosis, or apoptosis of hepatocytes, even developing into liver fibrosis and cirrhosis.33, 34, 35 Thus, its increase might also indicate a worsened prognosis of SFTS patients receiving IVIG therapy.

The current results must be interpreted with certain limitations owing to the study design. IVIG was not administered to patients in a randomized manner and despite of using PSM to produce well matched comparison groups for age, sex, delay of hospital admission and disease severity, there is potential for confounding within the results. IVIG was administrated to a subset of patients who received more other therapy, which might represent inconsistency that cannot be controlled by matching. To mitigate this bias, we further divided the patients into ribavirin receiving and non-receiving group, which again revealed no advantage of using IVIG. To reduce the impact of treatment preference among different physicians and in the different periods, we further analyzed the CFR by subgrouping the attending physicians and the hospitalization years. Consistent differences were observed across subgroups, suggesting the robustness of the conclusion (Supplementary Tables S14 and 15). Besides, the sample size limited the number of variables that could be included in the matching model, so we carefully prespecified a list of the most important prognostic factors. The dose of IVIG may have a significant impact on the results, which was not evaluated in the current study, since a standardized dose was applied in the hospital. Finally, the single-center study design might limit the extrapolation of the conclusion.

Despite of these limitations, the study had obvious advantages compared to previous case report studies, including the recruit of largest patient cohort in a single medical facility available, with the supportive therapy regimen kept in a comparable manner. The results provide persuasive data that increase the degree of certainty that IVIG offer no obvious efficacy in saving life or improving outcome of SFTS. Caution is needed for clinical physicians to continue prescribing IVIG, particularly when other efficacious treatments are available for SFTS.

Contributors

Q-B.L. and W.L. provided the conception of the study. J.D., N.C., X.Y., L.Z., W-X.Z., M.Y., Y-X.W., T.Y., X-A.Z., Z-D.Y. and H-D.L. collected the epidemiological data and conducted laboratory tests. S-S.Z., J.D., Q-B.L. and W.L. cleaned, analyzed, and interpreted the data. S-S.Z., J.D., Q-B.L. and W.L. drafted the manuscript. Q-B.L. and W.L. provided critical revision of the article for important intellectual content. The first two authors, S-S.Z. and J.D, as well as the corresponding authors Q-B.L. and W.L. accessed and verified underlying data. All authors read and approved the final version.

Data sharing statement

The study design, protocol and statistical analysis are provided in the main manuscript and the supplementary data files. The access to the data generated and analyzed in this study will be provided upon reasonable request to the corresponding author.

Declaration of interests

The authors have no competing interests to declare.

Appendix ASupplementary data

Supplementary Fig. S1.

Kinetics of key laboratory parameters from symptom onset between the Non-IVIG and IVIG groups. A: WBC, white blood cell; B: PLT, platelet; C: AST, aspartate transaminase; D: ALT, alanine transaminase; E: CK, creatine kinase; F: LDH, lactate dehydrogenase; G: ALB, albumin. Datapoints are median values and error bars show inter-quartile range. P values were calculated by generalized estimating equation. IVIG, intravenous immunoglobulin.