📄 other

Epidemiological Characteristics of HTLV Infection and Evaluation of ELISA Screening Performance Among Hospital-Based Populations in Ningde, China

This study characterizes the high prevalence of HTLV infection among hospital-based populations in Ningde, China, and evaluates the performance of ELISA screening, revealing a 1.30% initial reactive rate with 91.49% sensitivity and 95.19% specificity, while highlighting the need for PCR confirmation and optimized cut-offs in high-endemic settings.

Original authors: Wenqiang Wang, Zuodong Ye, Fenrui Fu, Yuying Qi, Renqing Ye

Published 2026-07-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Wenqiang Wang, Zuodong Ye, Fenrui Fu, Yuying Qi, Renqing Ye

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Invisible Passenger and the Sniffer Dog

Imagine your body is a bustling city, and sometimes, invisible passengers hitch a ride on your cells. One of these passengers is a virus called HTLV (Human T-cell Lymphotropic Virus). Unlike a flu bug that comes and goes quickly, HTLV is a "long-term tenant." Once it moves in, it stays for life, often hiding quietly without causing any trouble. But sometimes, this silent guest can wake up and cause serious problems, like a specific type of blood cancer or a condition that makes walking difficult. Because it hides so well, the only way to know if it's there is to test for it.

Doctors usually use a test called ELISA as their first line of defense. Think of ELISA as a highly sensitive sniffer dog. It's great at sniffing out the virus, but because it's so eager to find something, it sometimes barks at things that aren't the virus (false alarms) or, less often, misses a quiet guest (false negatives). The big question for scientists is: How good is this sniffer dog really, and does it work the same way for everyone? This is especially tricky in places where the virus is common, like Ningde City in China, where the "passengers" might be hiding in more people than we thought. If the dog barks too much, people get worried for no reason; if it doesn't bark enough, people miss out on help.

The Great Hospital Hunt

In this study, researchers from Ningde Municipal Hospital decided to put the ELISA sniffer dog through a serious workout. They didn't just look at healthy people; they looked at a massive crowd of 11,788 individuals, including people visiting the hospital because they were sick (outpatients and inpatients) and people just there for a routine check-up (physical examination participants). The goal was to see how often the virus showed up in these different groups and to see if the ELISA test was telling the truth.

To make sure the sniffer dog wasn't lying, the researchers used a "truth machine" called PCR (a type of genetic test) to double-check the results. Here's the clever part: usually, scientists only double-check the samples that the sniffer dog barked at. But these researchers knew that might miss the quiet guests. So, they also picked a special group of samples that the dog didn't bark at—specifically, those that were just barely below the barking line—to see if they were actually infected. This helped them get a much clearer picture of the dog's true accuracy.

What They Found

The hunt revealed some interesting patterns. First, the virus was definitely more common in Ningde than in many other places. Out of the 11,788 people tested, the ELISA test flagged 153 of them as potential carriers, which is a rate of 1.30%. But the story wasn't the same for everyone:

The Sick vs. The Healthy: People who were already in the hospital for other reasons were much more likely to have the virus (1.60%) compared to the healthy people just getting check-ups (0.46%).
Who is at Risk? The study found that if you were female, older, or currently a patient in the hospital, you were more likely to be carrying the virus. Being a patient made you about 2.5 times more likely to have it than someone just getting a check-up.

When they checked the sniffer dog's performance against the "truth machine" (PCR), the results were mixed but informative:

Overall Accuracy: The ELISA test was pretty good, catching about 91.49% of the real cases and correctly saying "no virus" to 95.19% of the virus-free people.
The Hospital Crowd: For the sick patients, the test was very good at finding the virus (95.08% sensitivity), but it barked a little too often at healthy people (lower specificity), meaning there was a higher chance of a false alarm.
The Check-Up Crowd: For the healthy people, the test was very good at saying "no virus" (99.57% specificity), but it missed more than a third of the actual cases (only 68.42% sensitivity). This means in a low-risk group, the dog might be too sleepy to catch the quiet guests.

The Golden Threshold

The researchers also looked at the "barking threshold"—the specific signal level that tells the dog to bark. The standard rule says the dog should bark if the signal is 1.0 or higher. However, the data suggested that in this high-risk area, the dog should be much more sensitive. They found that the dog should actually start barking at a signal as low as 0.3164 to catch the most cases without getting too many false alarms.

The Takeaway

The study concludes that HTLV is a real and relatively common issue in Ningde, especially among people who are already sick or older. While the ELISA test is a useful tool, it's not perfect for every situation. In hospitals, it might give a few false alarms, but in healthy check-up groups, it might miss the virus entirely. The researchers suggest that doctors in high-risk areas should consider lowering the "barking threshold" to catch more cases and always use the "truth machine" (PCR) to confirm any positive results. It's a reminder that one size doesn't fit all when it comes to catching invisible passengers.

Technical Summary: Epidemiological Characteristics of HTLV Infection and Evaluation of ELISA Screening Performance in Ningde, China

Problem Statement
Human T-cell lymphotropic virus (HTLV) infection is a significant public health concern in China, with Ningde City in Fujian Province identified as the region with the highest reported prevalence. However, existing epidemiological data and diagnostic performance evaluations are largely derived from blood donor cohorts or studies focusing on suspected infected individuals. These populations differ demographically and clinically from general hospital-based populations (outpatients, inpatients, and physical examination participants), limiting the generalizability of current findings to routine clinical screening. Furthermore, previous studies often suffer from "partial verification bias," where only ELISA-reactive samples are confirmed by PCR, potentially distorting sensitivity and specificity estimates. There is a critical need to characterize HTLV epidemiology in general hospital populations in this high-endemic region and to rigorously evaluate the diagnostic accuracy of ELISA screening under these specific conditions.

Methodology
This retrospective, observational study was conducted at Ningde Municipal Hospital between 2019 and 2024, enrolling 11,788 participants comprising outpatients, inpatients, and physical examination (PE) participants.

Screening Protocol: All samples underwent initial screening using a double-antigen sandwich ELISA kit (Beijing Wantai). Reactive samples were retested in duplicate; persistent reactivity defined an ELISA-reactive result.
Confirmation Strategy: To mitigate partial verification bias, the study employed a two-pronged confirmation strategy using real-time quantitative PCR (qPCR) as the reference standard:
1. All 153 ELISA-reactive samples were confirmed by PCR.
2. A non-random, preferentially selected subset of 487 ELISA-non-reactive samples was confirmed by PCR. This selection prioritized samples with higher Signal-to-Cutoff (S/CO) ratios (approaching the manufacturer's cut-off of 1.0) to maximize the detection of potential false negatives.
Statistical Analysis: Diagnostic performance metrics (sensitivity, specificity, PPV, NPV) were calculated. Binary logistic regression identified independent risk factors for seropositivity. Receiver Operating Characteristic (ROC) curve analysis was used to determine optimal S/CO cut-off values for different population subgroups.

Key Contributions
The study makes three primary methodological and empirical contributions:

Population Scope: It is among the first studies to concurrently evaluate ELISA screening performance across general outpatient, inpatient, and physical examination populations within a high-endemic region of China.
Bias Mitigation: By confirming a subset of ELISA-non-reactive samples (specifically those near the cut-off) alongside all reactive samples, the study addresses the partial verification bias common in prior blood-donor-based HTLV studies.
Threshold Optimization: The study derives population-specific optimal ELISA cut-off values through ROC analysis, moving beyond the manufacturer's default threshold to provide actionable screening parameters for clinical practice in high-prevalence settings.

Results

Epidemiology: The overall ELISA initial reactive rate was 1.30% (153/11,788). The reactive rate was significantly higher in the outpatient/inpatient group (1.60%) compared to the PE group (0.46%) ( $P < 0.001$ ).
Risk Factors: Multivariable logistic regression identified three independent risk factors for HTLV seropositivity:
- Outpatient/inpatient status (OR = 2.51).
- Female sex (OR = 1.60).
- Older age (OR = 1.03 per year increase).
Diagnostic Performance: Using PCR as the reference standard, the overall ELISA performance showed a sensitivity of 91.49% and specificity of 95.19%.
- Subgroup Variance: The outpatient/inpatient group exhibited higher sensitivity (95.08%) but lower specificity (91.42%), indicating a risk of false positives likely due to immune dysregulation in clinical patients. Conversely, the PE group showed markedly lower sensitivity (68.42%) despite high specificity (99.57%), suggesting a risk of missed diagnoses in this lower-prevalence, healthier cohort.
Optimal Cut-offs: ROC analysis identified an optimal S/CO cut-off of 0.3164 for the overall and outpatient/inpatient populations, and 0.3506 for the PE group. These values are substantially lower than the manufacturer's recommended threshold of 1.0.

Significance and Claims
The paper asserts that HTLV infection prevalence in Ningde is relatively high, particularly among clinical patients, suggesting that estimates based solely on blood donor cohorts may underestimate the true disease burden. The study concludes that while ELISA demonstrates acceptable overall screening performance, diagnostic accuracy varies significantly by population subgroup.

The authors claim that the findings underscore the necessity of:

Confirmatory Testing: The use of PCR or other confirmatory assays is essential, especially given the subgroup-specific risks of false positives (in clinical patients) and false negatives (in PE participants).
Population-Specific Optimization: In high-endemic regions like Ningde, lowering the ELISA cut-off threshold (e.g., to ~0.32) may improve sensitivity while maintaining acceptable specificity, contrasting with strategies in low-endemic regions that raise thresholds to reduce false positives.
Targeted Screening: The identification of female sex, older age, and clinical status as independent risk factors supports the implementation of targeted screening strategies in high-risk hospital populations.

The authors maintain a modest tone regarding the study's limitations, acknowledging that the single-center, cross-sectional design limits generalizability and that the enrichment strategy for non-reactive samples may have led to a slight underestimation of true sensitivity and NPV. They also note that the discordance between the ELISA (detecting HTLV-1/2) and PCR (detecting HTLV-1 only) could theoretically affect specificity, though they deem the impact minimal given the rarity of HTLV-2 in China.