Association between extreme temperature events and dengue risks in Dhaka City, Bangladesh

This study utilizes Distributed Lag Non-linear Models to analyze weekly dengue data in Dhaka City from 2016 to 2024, revealing that while maximum temperatures between 31.5°C and 33.2°C significantly elevate dengue risk, an increase in weekly heatwave days unexpectedly correlates with a protective effect against infection.

The Gist

🌡️ The Big Picture: Dengue, Dhaka, and the Weather

Imagine Dhaka, Bangladesh, as a giant, bustling kitchen. For years, this kitchen has been getting hotter and more crowded, and unfortunately, it's become the perfect breeding ground for Dengue fever (a nasty virus carried by mosquitoes).

The researchers wanted to answer a simple question: How does the weather change the number of dengue cases? specifically, they looked at two things:

1. Maximum Temperature: How hot it gets during the day.
2. Heatwaves: Days where it gets really hot for several days in a row.

They used a fancy statistical tool called a DLNM (Distributed Lag Non-Linear Model). Think of this tool as a time-traveling magnifying glass. It doesn't just look at the weather today; it looks at how the weather from the last 6 weeks affects the number of sick people right now.

---

🔥 Finding #1: The "Goldilocks" Zone for Mosquitoes

The study found that the relationship between heat and dengue isn't a straight line. It's more like a rollercoaster or a hill.

• The Sweet Spot (31.5°C – 33.2°C): This is the "Goldilocks" zone. It's not too cold, and it's not too hot. In this temperature range, the dengue risk is at its highest.

  • The Analogy: Imagine the mosquitoes are like bakers. When the kitchen is at a perfect warm temperature (around 32°C), they are super productive. They bake eggs faster, bite people more often, and the virus inside them grows quickly. It's like a factory running at 100% efficiency.
  • The Result: If the temperature is in this range, the risk of getting dengue goes up by about 18% compared to average days.

• The Danger Zone (> 35.5°C): Once the temperature gets extremely hot (above 35.5°C), the risk actually drops.

  • The Analogy: If the kitchen gets so hot that the ovens are on fire, the bakers (mosquitoes) get exhausted. They stop working, they die off, or they just can't function. The virus inside them also struggles to survive. It's like a factory that has to shut down because the heat is too intense for the workers.
  • The Result: Extreme heat acts as a natural "off switch" for the outbreak.

🌪️ Finding #2: The Heatwave "Shield"

The researchers also looked at Heatwaves (defined as days where it's very hot for at least two days in a row).

• The Surprise: You might think more heatwaves mean more disease. But the study found the opposite! Having 3 days of heatwaves in a single week actually reduced the risk of dengue by about 72%.
  • The Analogy: Think of a heatwave as a heat blast that sweeps through the city. It's like a giant, invisible broom that sweeps the mosquitoes out of the neighborhood. For a short time, the mosquitoes are too stressed to bite or reproduce.
  • The Catch: This is a short-term effect. If the heatwave lasts too long or happens at the wrong time, the mosquitoes might eventually come back, but for the first few weeks, the city is safer.

🌧️ What About Rain and Humidity?

You might expect rain to be the main culprit (since mosquitoes love puddles). However, in Dhaka, the study found that rain and humidity didn't have a huge impact on the short-term risk.

• Why? Dhaka is a concrete jungle. Even when it doesn't rain, people store water in buckets, tanks, and plastic containers. The mosquitoes have plenty of "puddles" to live in all year round, so a little extra rain doesn't change the game much.

🧬 The Changing Cast of Characters (Serotypes)

Dengue isn't just one virus; it has different "flavors" called serotypes (like DENV-2 and DENV-3). The study noticed that the "main character" in Dhaka changed over the years.

• Even though the specific virus changed, the temperature rules stayed the same. Whether it was the 2019 virus or the 2023 virus, they all loved the 31.5°C–33.2°C range and hated the extreme heat.

🚨 Why Does This Matter? (The "So What?")

This research is like a weather forecast for disease.

1. Early Warning Systems: If the weather forecast predicts a week where temperatures will hover around 32°C, health officials can sound the alarm before the outbreak happens. They can spray for mosquitoes and tell people to cover their water tanks immediately.
2. Climate Change Warning: As the world gets warmer, Dhaka might spend more time in that dangerous "Goldilocks" zone. This means we could see more frequent and larger dengue outbreaks.
3. The Silver Lining: If the climate gets extremely hot (above 35.5°C), it might naturally suppress the virus, but that's a dangerous trade-off because extreme heat kills people in other ways (heatstroke, dehydration). We can't rely on extreme heat to save us.

🏁 The Bottom Line

• Moderate Heat (31-33°C): DANGER. Mosquitoes are happy, biting, and spreading the virus fast.
• Extreme Heat (>35°C): SAFE (for now). The heat is too much for the mosquitoes, and the risk drops.
• Action Plan: Policymakers need to use this data to build a "Dengue Dashboard." When the temperature hits that sweet spot, it's time to fight back hard with mosquito control.

In short: Don't wait for the rain to worry about dengue in Dhaka; watch the thermometer. If it's getting warm (but not scorching), the mosquitoes are coming.

Technical Summary

1. Problem Statement

Dengue fever has become a critical public health crisis in Bangladesh, with case numbers increasing eighteen-fold between 2000 and 2024. Dhaka City serves as the epicenter of these outbreaks. While the link between climate and dengue is established, existing research in Bangladesh has limitations:

• Methodological Gaps: Previous studies often relied on linear regression models or fixed lag structures, failing to capture the complex, non-linear, and delayed (distributed lag) relationships between weather variables and disease incidence.
• Contextual Gaps: There is a lack of high-resolution (weekly) data analysis specifically for Dhaka City regarding heatwaves and maximum temperature extremes.
• Knowledge Gap: The specific impact of sustained extreme heat (heatwaves) on dengue transmission in a hyperendemic urban setting remains poorly understood, with conflicting evidence globally regarding whether heat suppresses or exacerbates transmission.

2. Methodology

The study employed a robust statistical framework to analyze weekly data from 2016 to 2024 in Dhaka City.

• Data Sources:
  • Dengue Cases: Daily reported cases from the Directorate General of Health Services (DGHS), aggregated into weekly counts.
  • Meteorological Data: Daily maximum temperature, relative humidity, and rainfall from the Bangladesh Meteorological Department (BMD).
  • Covariates: Population data (World Bank), hospital closure days (public/medical holidays), and dominant serotype shifts (DENV-2 vs. DENV-3).
• Statistical Model:
  • Distributed Lag Non-linear Model (DLNM): Implemented using the R package dlnm. This method simultaneously models non-linear exposure-response relationships and delayed effects (lags) over a 7-week period (0–6 weeks).
  • Model Type: Two separate Negative Binomial Generalized Linear Models (GLMs) were constructed to handle overdispersion in count data:
    1. MaxT Model: Analyzing weekly maximum temperature.
    2. Heatwave (HW) Model: Analyzing the count of heatwave days per week.
  • Heatwave Definition: A heatwave day was defined as a day where the maximum temperature exceeded the 90th percentile (35.5°C) for at least two consecutive days.
  • Confounding Control: Models adjusted for relative humidity, cumulative rainfall, seasonality (natural cubic splines), month of the year, hospital closures, and population offsets.
• Validation:
  • Sensitivity analyses were conducted by varying degrees of freedom (DOF) for time trends and spline functions.
  • Sub-period analyses were performed to assess associations during different dominant serotype periods (2016–2018: DENV-2; 2019–2022: DENV-3; 2023–2024: DENV-2).

3. Key Contributions

• High-Resolution Temporal Analysis: This is one of the first studies in Bangladesh to utilize weekly data and DLNM to dissect short-term (0–6 weeks) lagged effects of temperature extremes on dengue.
• Heatwave Specificity: It provides novel evidence on the specific impact of heatwave duration (number of days per week) on dengue risk, distinguishing between moderate and extreme heat exposure.
• Serotype-Stratified Insights: The study investigates how these climatic associations vary across different dominant dengue serotypes, offering a more nuanced understanding of vector-virus-climate interactions.
• Policy-Ready Metrics: The study translates complex statistical outputs into actionable Relative Risk (RR) estimates for specific temperature thresholds and heatwave frequencies.

4. Key Results

A. Maximum Temperature (MaxT) Association

• Pattern: An inverted U-shaped (parabolic) relationship was observed between maximum temperature and dengue risk.
• Peak Risk: The highest risk occurs in the range of 31.5°C to 33.2°C.
  • At 33.2°C, the Relative Risk (RR) peaked at 1.186 (95% CI: 1.002–1.403) compared to the mean reference (31.5°C).
  • A 1°C increase from the mean resulted in a 15% increase in risk.
• Protective Effect of Extreme Heat: Temperatures exceeding 35.5°C showed a protective effect (RR < 1). At 38°C, the risk dropped significantly (RR: 0.191), suggesting that extreme heat inhibits mosquito survival or virus replication in the short term.

B. Heatwave Days Association

• Pattern: A reverse J-shaped (protective) association was found between the number of heatwave days per week and dengue risk.
• Optimal Protection: The lowest risk was observed at 3 heatwave days per week, with an RR of 0.275 (95% CI: 0.178–0.423), representing a 72.5% reduction in risk compared to weeks with no heatwaves.
• Non-linearity: While moderate heatwaves (1–4 days) were protective, the risk began to rise slightly at 6–7 days, though confidence intervals widened due to data scarcity at these extreme frequencies.

C. Sensitivity and Sub-Period Analysis

• Robustness: Results remained consistent across different model specifications (varying DOF) and heatwave definitions.
• Serotype Influence:
  • DENV-2 (2016–2018 & 2023–2024): Showed the classic parabolic risk curve.
  • DENV-3 (2019–2022): Also showed a parabolic curve with peak risk around 33.8°C.
  • Heatwave Effect: The protective effect of heatwaves (3–5 days) was consistent across all sub-periods and serotypes.

D. Other Climatic Factors

• Rainfall & Humidity: No statistically significant short-term association was found between cumulative rainfall or relative humidity and dengue risk in the final models, likely due to the year-round breeding potential in Dhaka's urban environment (water storage containers).

5. Significance and Implications

• Early Warning Systems (EWS): The findings support the development of a "Dengue Risk Dashboard" for Dhaka. Policymakers can use the identified temperature threshold (31.5–33.2°C) as a trigger for intensified vector control interventions.
• Climate Change Adaptation: As Bangladesh faces rising temperatures, the study suggests a dual scenario:
  1. Increased Risk: If temperatures shift into the 31.5–33.2°C "danger zone," dengue cases will surge.
  2. Suppression: If temperatures exceed 35.5°C (extreme heatwaves), transmission may be temporarily suppressed, though this comes with other heat-related health risks.
• Strategic Planning: The results align with the Bangladesh National Dengue Prevention and Control Strategy (2024–2030), emphasizing climate-sensitive surveillance. The study highlights that vector control must be timed specifically for the "optimal temperature window" rather than just the rainy season.
• Limitations: The study relies on surveillance data which may underreport mild/asymptomatic cases and does not account for vector density or specific urban infrastructure variables. However, the use of syndromic surveillance data provides a practical proxy for transmission dynamics in hyperendemic regions.

Conclusion: The study establishes that while moderate warming (up to ~33°C) significantly accelerates dengue transmission in Dhaka, sustained extreme heat (heatwaves) acts as a short-term suppressor. This non-linear, lagged relationship is critical for designing evidence-based, climate-resilient public health interventions.