21JUL

One in six infections beats the drug

4 min read

09:12UTC

WHO's GLASS 2025 report found one in six of 23 million bacterial infections across 104 countries no longer responds to first-line antibiotics, with resistance highest where labs are scarcest.

← Back to Ebola outruns its own contact tracing Jump to analysis ↓

ScienceDeveloping

Key takeaway

Drug resistance is measured worst where it is highest, so the one-in-six global figure is a floor, not a ceiling.

The World Health Organization (WHO) published its 2025 Global Antimicrobial Resistance Surveillance System (GLASS) report drawing on 23 million laboratory-confirmed bacterial infections across 104 countries in 2023 ¹. GLASS pools resistance data from national laboratories to track global trends. The headline finding: one in six of those infections no longer responds to the antibiotic a doctor would reach for first, and resistance rose in 40% of the pathogen-drug combinations WHO tracks between 2018 and 2023.

In WHO's South-East Asia and Eastern Mediterranean regions, one in three common infections is now drug-resistant; in Europe the rate is one in ten, in Africa one in five ². For an individual patient, resistance means the first prescription fails more often, pushing treatment onto second-line drugs that are costlier, more toxic and scarcer in exactly the settings already struggling.

The measurement itself carries the warning. GLASS counts only laboratory-confirmed infections, so its figures are anchored to wherever lab capacity exists. Because resistance is highest where surveillance is thinnest, the true rate in under-resourced systems is systematically undercounted, which means the report understates the burden rather than overstating it. This is the slow pandemic the Global Action Plan on Antimicrobial Resistance adopted at the World Health Assembly last month is meant to bend, targeting a 10% cut in resistance deaths by 2030 , and the one the Food and Agriculture Organization has costed at $318bn in livestock losses by 2040 if left unchecked .

Deep Analysis

In plain English

Antibiotics are the drugs that kill bacterial infections, everything from a chest infection to sepsis after surgery. Antimicrobial resistance (AMR) means the bacteria have evolved to survive the drugs designed to kill them. In 2023, one in six bacterial infections tracked by the World Health Organization across 104 countries had already become resistant to the standard first-line antibiotic prescribed to treat it. Resistance is growing. Some infections that used to respond to cheap, widely available drugs now need more expensive second-line treatments, or do not respond well to any available drug. GLASS 2025 data puts the South-East Asian and Middle Eastern rate at one in three bacterial infections already resistant. The WHO report is a measurement of a problem that has been building for decades and is getting worse.

Deep Analysis

Root Causes

The 1 in 6 global resistance rate reflects two structurally distinct drivers that the GLASS report separates geographically. In high-income countries, resistance rises primarily from inappropriate prescribing in outpatient settings: antibiotics prescribed for viral infections, failure to complete courses, and livestock prophylactic use.

EARS-Net data for Europe shows resistance climbing despite a decade of stewardship programmes, because prescribing reform has been offset by acquisition in healthcare settings. In low- and middle-income countries, the dominant driver is inadequate sanitation: populations that lack clean water develop resistance through sustained low-level environmental antibiotic exposure from agricultural and pharmaceutical runoff, not primarily from clinical prescribing.

What could happen next?

Risk
If resistance increased in 40% of pathogen-drug combinations between 2018 and 2023, and the GAP-AMR 2030 target requires only a 10% reduction in AMR deaths, the target may be achievable by statistical baseline adjustment rather than actual resistance reversal.
Meaning
GLASS 2025 confirms that resistance is highest where laboratory capacity is scarcest, creating a recursive measurement problem: countries with the worst burden generate the least data, meaning GLASS systematically underestimates the global burden.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: The 2013 CDC AMR Threat Report introduced the first formal AMR threat categorisation in the US and triggered the Obama administration's 2014 National Strategy for Combating Antibiotic-Resistant Bacteria, which committed $1.2bn over five years and established CARB-X (the Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator) for early-stage antibiotic R&D.

The GLASS 2025 report performs the same function globally that the 2013 US report performed nationally: it creates a quantified baseline against which policy commitments can be held. The difference is that the CDC's 2013 threat report arrived before the US government had committed to a target; the GLASS 2025 data arrives after WHA79 has already adopted a framework with a 2030 deadline.

Counter-parallel: The 2016 Review on Antimicrobial Resistance (the Jim O'Neill review, commissioned by the UK government) projected 10 million AMR deaths per year by 2050, a figure that shaped the G7/G20 AMR frameworks of the subsequent decade. O'Neill's modelling was criticised at the time for overestimating resistance spread and underestimating substitution effects.

The GLASS 2025 empirical data on 23 million actual infections from 104 countries provides a ground-truth check on those projections; the trajectory (40% of combinations rising in five years) is consistent with the O'Neill directional warnings, even if the annual death figures remain contested.

FAO costed AMR inaction in livestock at $318bn by 2040 , roughly six times the cost of acting now.

The human health bill compounds that figure: the GLASS 2025 data implies that if 40% of drug-combination resistance is rising across a baseline of 23 million confirmed infections, with the true global burden likely two to three times higher when Community-based and informal-sector infections are included, the economic losses from treatment failure, extended hospitalisation, and productivity loss run to several hundred billion dollars annually.

The World Bank's 2017 AMR economic modelling projected GDP losses of 1.1-3.8% in low-income countries by 2050 if resistance trends are not reversed. The GLASS 2025 trajectory is consistent with the scenario under which those projections materialise.

Consensus view: The GLASS 2025 report is the most comprehensive resistance dataset assembled to date, and the Global Antibiotic Research and Development Partnership (GARDP) and the Global AMR R&D Hub both assess its headline finding, resistance increasing in 40% of pathogen-drug combinations between 2018 and 2023, as consistent with national-level surveillance data from EARS-Net (European AMR surveillance).

The key analytical advance in GLASS 2025 is geographic granularity: it can now show that resistance in South-East Asia and the Eastern Mediterranean is three times the European rate, which ties the burden directly to diagnostic capacity. Countries that cannot distinguish resistant from susceptible infections do not report them, so the true global figure is higher than GLASS measures.

Counter-view: India's Department of Biotechnology and South Africa's National Institute for Communicable Diseases (NICD) have both noted that the GLASS dataset systematically under-represents primary care, where most antibiotic use and resistance selection occurs, because participating laboratories are overwhelmingly hospital-based.

The 23 million infections sampled represent the most serious cases. Resistance in Community-acquired infections, typically treated empirically without laboratory confirmation, is invisible to GLASS. The 1 in 6 global figure is therefore a floor on the true resistance burden, not a central estimate.

Key tension: Whether the GAP-AMR framework adopted at WHA79 , targeting a 10% cut in AMR deaths by 2030, is calibrated to the measured GLASS burden (which is itself a floor) or to the true burden, and whether the 10% target is achievable or cosmetically modest given the trajectory the 2018-2023 trend line reveals.

Sources:CIDRAP / WHO·CIDRAP / WHO

Mentions:World Health Organization →WHO GLASS →FAO →East Asia →WHA79 →Mediterranean →India →South Africa →GlassWorm →CDC →Global Action Plan on AMR →CIDRAP →

First Reported In

Update #8 · Ebola reaches France through a screening blind spot

CIDRAP / WHO· 25 Jun 2026

Read original →

Causes and effects

Caused by

WHA79 adopts a 10-year AMR plan

The GLASS 2025 data release provides the empirical baseline against which the GAP-AMR framework adopted at WHA79 must now be measured.

Occurred 23 May 2026

Read story →

FAO costs drug-resistance inaction at $318bn

The FAO's $318bn livestock-loss costing contextualises the economic stakes against which the GLASS 2025 rising-resistance data lands.

Occurred 3 Jun 2026

Read story →

This Event

One in six infections beats the drug

Resistance runs highest exactly where laboratories are fewest, so the published one-in-six figure is a floor rather than the true global rate.

Led to

WHO issues its first fungal blueprint

WHO GLASS's bacterial-only resistance findings left fungal disease outside AMR surveillance until this Blueprint.

Occurred 1 Jul 2026

Read story →

Different Perspectives

UK Government

The Biological Security Strategy implementation report published 14 July commits the UK to deciding within twelve months whether to make DNA-synthesis screening mandatory, backed by GBP 1.83bn for biosecurity centres. It is a preemptive move on a threat that has not yet produced a UK incident.

Cambodia's Ministry of Health

Cambodia confirmed a nine-month-old girl in Phnom Penh as its fourth 2026 H5N1 case on 11 July, the country supplying four of the world's seven new human cases since February. The ministry's confirmed-case cadence is running well ahead of a US surveillance system that has recorded nothing new since March.

Knowledge Ecology International

KEI published proposed annex text on the day IGWG7 opened that would attach benefit-sharing duties to a product's market registration, regardless of whether the manufacturer used shared samples or only downloaded sequence data. It is the Africa Group's fallback position if the sample-based bargain fails to bind.

European Union

The EU and other industrialised states backed a hybrid model built on centralised infrastructure with more flexible access terms, the opposite design choice from the Africa Group's. Brussels wants conditions handled contractually once samples reach shared facilities, not attached to national custody of the material itself.

Africa Group / Algeria

Algeria, speaking for the Africa Group at IGWG7, backed distributed national and regional laboratories under binding access conditions, arguing that supplying states must keep legal control of samples rather than surrender them into shared infrastructure.

Uganda's Ministry of Health

Uganda discharged its last confirmed patient on 16 July and opened a 42-day countdown to declaring its own outbreak over, closing at 20 cases and 2 deaths with no new case since 21 June. This is a closing chapter running alongside DR Congo's still-worsening one, contingent on a border DRC can no longer see across.