Long-Term Fasting Rewires Your Gut’s Viral Ecosystem in Surprising Ways

When researchers talk about fasting and the gut microbiome, they almost always mean bacteria. But bacteria do not act alone: bacteriophages, viruses that infect and kill bacteria, regulate which bacterial species thrive and which collapse. A new study published in NPJ Biofilms and Microbiomes by Falshaw N et al. is the first to systematically map how the gut virome responds to long-term fasting, and the picture that emerges is more dynamic, and more persistent, than anyone expected.

Fasting Shrinks Gut Virome Diversity and Shifts Phage Balance

Across 89 participants who fasted for an average of 9.8 days at roughly 250 calories per day, the gut virome changed in measurable and consistent ways. Alpha diversity, measured by both Shannon index and species richness, fell significantly by the end of the fast. The overall community structure also shifted, with the most pronounced separation appearing between baseline samples and those collected at the end of the fast.

More striking was what happened to phage lifestyle. Bacteriophages broadly fall into two categories: virulent phages that immediately lyse and kill their bacterial hosts, and temperate phages that integrate quietly into bacterial genomes as prophages. Fasting pushed the ratio toward virulent phages, not because virulent richness increased, but because temperate phage diversity declined. In other words, the quiet, integrated phages disappeared disproportionately, leaving a virome weighted toward active bacterial predation.

Out of 10,863 viral species detected across all samples, crAss-like viruses were the most prevalent, appearing in 78.8% of participants and reaching nearly 40% of the virome in some individuals. The most common annotated families were Siphoviridae, Myoviridae, Podoviridae, and Microviridae. Notably, 43% of detected viral taxa could not be assigned to any known family at all, a reminder that characterizing viral communities remains technically harder than profiling bacteria or archaea because viruses lack the conserved genetic markers that make bacterial surveys straightforward.

Why These Viral Shifts Could Matter for Metabolic Health

Bacteriophages are not bystanders. As research on the gut virome and human health has established, phages regulate bacterial population sizes, transfer genes horizontally, and shape the metabolic outputs that link microbiome composition to host physiology. During fasting, the bacteria that disappear are largely the ones that ferment dietary fiber into short-chain fatty acids like butyrate. If phages are contributing to that bacterial collapse, they could be amplifying or accelerating the metabolic reset that fasting produces.

The study found that 49 consensus viral species showed consistent directional shifts across the discovery cohort and two independent validation groups. Thirty of those 49 could be mapped to predicted bacterial hosts, and most showed concordant abundance changes: when a bacterial genus declined, its associated phages declined with it. Faecalibacterium phages, for instance, dropped during fasting, a pattern that contrasts sharply with what is seen in inflammatory bowel disease, where the same phages are enriched. That divergence hints that phage-host dynamics under nutrient stress operate differently from those under inflammatory conditions, a distinction that could matter for how fasting-based interventions are evaluated clinically.

One phage stood out as particularly interesting: Mushuvirus, linked to Faecalibacterium and Ruminococcaceae. It emerged as a highly connected hub in the viral-bacterial network at the end of the fast and again at three-month follow-up. Mushuvirus encodes a diversity-generating retroelement that may enhance its adaptability and, potentially, mucin binding, suggesting it plays an active structural role in how the microbiome reorganizes after a fast.

How the Study Was Built to Find Reproducible Signals

The core dataset came from re-analyzed bulk gut metagenomic samples drawn from a medically supervised fasting protocol. Samples were collected at four timepoints: baseline, end of fast, one month after refeeding, and three months after refeeding. The team used the Phanta bioinformatics pipeline for virome composition, linear mixed-effects models and PERMANOVA for diversity statistics, and BACPHLIP to classify phage lifestyles as virulent or temperate.

To confirm that the 121 differentially abundant viral taxa were not cohort-specific noise, the researchers compared results against two independent validation fasting cohorts. The 49 consensus species that survived that cross-cohort filter represent the reproducible fasting virome signature. For network analysis, a subset of 28 participants with complete longitudinal data was used to build cross-domain co-abundance networks via FastSpar, mapping virus-bacterium relationships at the genus level across all four timepoints.

Virome Network Complexity Stays Altered for Months After Refeeding

The network findings are where this study gets genuinely unexpected. Gut virome diversity largely recovered to baseline by the three-month follow-up. But the underlying structure of the viral-bacterial interaction network did not. Cross-domain network complexity followed a biphasic pattern: it increased at the end of the fast, simplified sharply at one month, then increased again by three months. The negative correlations between viral and bacterial taxa dominated at every timepoint, accounting for 82 to 94% of all network edges, consistent with phages actively suppressing bacterial populations rather than co-existing neutrally.

Even at 90 days post-fast, the network architecture remained substantially different from baseline. The diversity numbers might look recovered, but the wiring between species had reorganized. The researchers describe this as a 'controlled and coordinated restructuring' of microbiome dynamics, distinct from the dysregulated trajectories seen in inflammatory disease. Highly connected bacterial hubs in the network included Akkermansia, Faecalibacterium, Solobacterium, and Faecalibacillus, species with known roles in gut barrier function and metabolic signaling.

Those challenges in virome research are real here too: the study relied on bulk metagenomics, which may undercount free viral particles, and the reference-based pipeline likely misses genuinely novel phages. Lifestyle classifications reflect genomic potential, not confirmed active lysis. Participant numbers shrank at follow-up timepoints, reducing statistical power. And dietary changes of any kind can measurably shift the gut virome, making it hard to isolate fasting-specific effects from the confounding influence of refeeding foods.

What Comes Next in Gut Virome and Fasting Research

The practical implication is that fasting protocols cannot be fully understood by tracking bacteria alone. The virome rewires in ways that persist well beyond the return of normal eating, and those rewired networks involve phages with known connections to metabolic health markers. Fasting already improved glycaemic control, lipid metabolism, and liver function in these participants. Whether the viral layer is contributing mechanistically to those outcomes, or simply reflecting bacterial shifts downstream, is the open question this study leaves squarely on the table.

Directly linking virome features to host metabolic phenotypes will require larger cohorts, longitudinal virome-specific sequencing, and better reference databases for the roughly 43% of viral taxa that currently have no family-level annotation. For now, the gut virome has moved from a largely unexplored footnote in fasting biology to a documented, reproducible, and structurally persistent part of the story.

Evidence rating

EMERGING — 80/100

Caveats:

• Single-arm observational intervention — no control group, so changes cannot be causally attributed to fasting alone.
• Metagenomics captures genomic potential, not direct evidence of active lytic phage replication or confirmed biological interactions.
• Network correlations (viral-bacterial co-abundances) are statistical associations, not proof of direct phage-host relationships.
• Smaller participant numbers at 1- and 3-month follow-up reduce statistical power for longitudinal claims.
• Funding was not disclosed by the authors.

Single-arm metagenomic intervention study (n=89, 241 samples) tracking the gut virome before, during, and up to three months after approximately 10 days of supervised very-low-calorie fasting. Findings show associations between fasting and virome restructuring, validated directionally in two independent cohorts, but the design cannot establish causation, active phage replication was not directly measured, and network analyses reflect statistical co-abundance rather than confirmed biological interactions. The virome-fasting relationship has not yet been examined in a randomized controlled trial.

References

1. Falshaw N et al. (2026). Remodelling of the gut virome after long-term fasting. NPJ biofilms and microbiomes. https://doi.org/10.1038/s41522-026-01069-6
2. Human Virome and Disease: High-Throughput Sequencing for Virus Discovery, Identification of Phage-Bacteria Dysbiosis and Development of Therapeutic Approaches with Emphasis on the Human Gut – MDPI. mdpi.com. https://www.mdpi.com/1999-4915/11/7/655
3. Rahul Harikumar Lathakumari, Leela Kakithakara Vajravelu, Anusha Gopinathan, et al. (2025). The gut virome and human health: From diversity to personalized medicine. Engineering Microbiology. https://doi.org/10.1016/j.engmic.2025.100191
4. Zhiyang Feng, Elke Burgermeister, Anna Philips, Tao Zuo, Weijie Wen (2025). The gut virome in association with the bacteriome in gastrointestinal diseases and beyond: roles, mechanisms, and clinical applications. Precision Clinical Medicine. https://doi.org/10.1093/pcmedi/pbaf010
5. Evan Pargin, Michael J. Roach, Amber Skye, et al. (2023). The human gut virome: composition, colonization, interactions, and impacts on human health. Frontiers in Microbiology. https://doi.org/10.3389/fmicb.2023.963173