The Indian Council of Medical Research’s antimicrobial resistance surveillance report for 2025 records resistance at levels that leave clinicians with very little to prescribe. Nine in ten Acinetobacter baumannii isolates resisted carbapenems, the class held in reserve for the most serious infections. Fewer than one in ten Escherichia coli isolates responded to ciprofloxacin. For the first time the network extended beyond large tertiary hospitals to 34 secondary-care hospitals, including 12 district hospitals — an important methodological step, because what happens outside metropolitan referral centres has until now been largely invisible.
Antimicrobial resistance is not a future risk in India. It is a present condition of its hospitals, measured annually, and the measurements are worsening. The Indian Council of Medical Research’s Antimicrobial Resistance Surveillance and Research Network, AMRSN, published its annual report for 2025 as World AMR Awareness Week opened on 18 November, and the figures describe a system in which the drugs of last resort are frequently no longer effective.
What the report covers
The network analysed 1,13,008 culture-positive samples from 30 large tertiary-care hospitals during 2025. Crucially, it also analysed 20,529 positive samples from 34 secondary-care hospitals, including 12 district hospitals — the first time the network has reported from that tier.
That extension deserves emphasis. Resistance data drawn only from large referral hospitals is systematically biased: those institutions receive the sickest patients, who have usually already been treated elsewhere, so their resistance rates are higher than the national picture and their case mix is unrepresentative. Extending surveillance to district hospitals begins to answer a question Indian policy has not been able to answer — what resistance looks like where most Indians actually receive care.
The organisms and the numbers
In the tertiary hospitals, four organisms dominated: Escherichia coli at 24.4 per cent of isolates, Klebsiella pneumoniae at 16.9 per cent, Pseudomonas aeruginosa at 12.2 per cent and Acinetobacter baumannii at 9.5 per cent. All four are Gram-negative bacteria, and all four are common causes of hospital-acquired pneumonia, bloodstream infection and urinary tract infection.
The susceptibility figures are where the report becomes alarming:
- E. coli: 18.4 per cent of isolates susceptible to cefotaxime, 24.7 per cent to ceftazidime, and 9 per cent to ciprofloxacin. A drug effective in fewer than one in ten cases is not a usable empirical choice.
- Klebsiella pneumoniae: only 36.5 to 40.4 per cent still susceptible to carbapenems, and lower in intensive-care samples.
- Acinetobacter baumannii: 91.1 per cent carbapenem-resistant.
- Staphylococcus aureus: 56.8 per cent methicillin-resistant, with some linezolid resistance also reported.
Carbapenems matter more than the other numbers. They are broad-spectrum beta-lactams held deliberately in reserve for serious Gram-negative infection, and for many such infections they are the last well-tolerated option. When fewer than two in five Klebsiella isolates respond and nine in ten Acinetobacter isolates do not, the clinician’s remaining choices are older, more toxic agents such as colistin, or combinations with thin evidence.
For enteric and diarrhoeal infections the figures are similarly stark. Ciprofloxacin susceptibility was 7.5 per cent in Shigella flexneri, 8.9 per cent in S. sonnei and 0.7 per cent in diarrhoeagenic E. coli; cefixime susceptibility in diarrhoeagenic E. coli was 4.6 per cent. Salmonella Typhi, the typhoid organism, still responds to azithromycin and ceftriaxone, but its ciprofloxacin susceptibility is described as extremely low. The practical message the report draws is that fluoroquinolones should not be used indiscriminately for diarrhoeal illness and that severe or persistent cases warrant laboratory testing.
The newborn data
The report’s separate analysis of neonates — infants up to 28 days — is the finding with the sharpest human edge. Klebsiella pneumoniae accounted for 25.1 per cent of bloodstream isolates in newborns, with Candida species and Acinetobacter baumannii also common, and susceptibility to third-generation cephalosporins was poor.
Neonatal sepsis is a condition in which hours matter. Treatment must begin before culture results are available, which means it must be based on what the local resistance pattern suggests will work. If a quarter of neonatal bloodstream infections are caused by an organism poorly susceptible to the standard empirical choice, then the standard choice is failing some of those infants before anyone knows it.
What the report recommends
The recommendations are unglamorous and largely about hospital practice: strengthen infection prevention and control, which means hand hygiene, sterilisation of instruments and cleaning of surfaces and equipment; make laboratory diagnostics reliable and available; guide treatment by susceptibility results rather than habit; protect carbapenems and other last-resort agents through disciplined prescribing; and build region-specific surveillance and guidelines, because resistance patterns differ between places.
There is a reason the emphasis falls on infection control rather than on new drugs. Resistant organisms in hospitals spread primarily by transmission between patients, via hands, instruments and surfaces. An infection prevented needs no antibiotic at all, which makes infection control the only intervention that reduces both resistance and its consequences simultaneously. It is also, in practice, the hardest to sustain, because it depends on thousands of repeated acts by overstretched staff rather than on a procurement decision.
Context, and the international echo
An earlier ICMR study, reported in September 2026, examined 26,213 confirmed Gram-negative infections across 20 hospitals between 2022 and 2025 and found 61.1 per cent carbapenem-resistant, with higher mortality where resistance was present — 31.2 per cent against 23.5 per cent for K. pneumoniae — and treatment costs one to two times higher. India launched its National Action Plan on Antimicrobial Resistance 2.0, covering 2025 to 2029, in November 2025.
The same week brought a comparable warning from Europe, where the European Centre for Disease Prevention and Control described the spread of carbapenem-resistant Enterobacterales and clinicians reported patients for whom no effective agent remained. This is not an Indian problem with an international footnote; it is a global problem on which India has unusually good surveillance data and unusually severe numbers.
Limitations
NSH was unable to locate the AMRSN 2025 report itself on icmr.gov.in. All the figures above come from a single news report of 18 November, corroborated only in general terms by a second outlet. A separate ICMR annual report for 2023 is publicly available and NSH consulted it for context, but it is a different document with different figures and the two should not be conflated. No ICMR official is quoted in the coverage, no year-on-year comparison with 2024 is given, and NSH found no ICMR press release. The figures should be checked against the report before republication.
– Rashmi M



