Health Risks Associated With a BMI Below 18.5: The Under-Discussed Low End

The scale dropped again and the app flashed 'underweight'. You are not trying to lose weight. You need to know what a BMI below 18.5 actually means for your health, beyond the label. Here is what the epidemiological evidence says about the risks tied to low body weight, and how much of that risk comes from the weight itself versus other factors.

Body mass index is screening arithmetic at population scale, not a diagnostic for an individual body. It divides weight in kilograms by height in metres squared and sorts adults into bands fixed by a World Health Organization committee in the 1990s using epidemiological cut-points, not physiological thresholds. The WHO adult underweight threshold of BMI < 18.5 kg/m² was published in WHO Technical Report Series 854 (1995) and reaffirmed in Obesity: Preventing and Managing the Global Epidemic (2000). The CDC, NHS, NICE, and NHLBI all use the same cut-off: a BMI below 18.5 is classified as underweight.

The mistake newcomers make is treating a BMI number as a direct readout of metabolic health or body fatness for a specific person. It was designed to spot statistical shifts in large groups. A BMI of 17.5 does not tell you if someone is malnourished, has sarcopenia, or is a naturally lean endurance athlete. It tells you their weight is low for their height. What follows describes what the data show about the risks associated with that low weight, and where the signal gets confused by pre-existing disease and smoking.

What the Global BMI Collaboration Found

The most cited evidence is the Global BMI Mortality Collaboration, a 2016 meta-analysis in The Lancet that pooled 10,625,411 participants across 239 prospective studies on four continents. It found a J-shaped relationship between BMI and death from any cause. The lowest death rates sat in the BMI range 22.5-25.0 kg/m². Rates climbed at both ends of the distribution.

For people with a BMI between 15.0 and 18.5, the hazard ratio for death from any cause was 1.47 (95% CI 1.39-1.55) for men and 1.33 (95% CI 1.26-1.41) for women, compared to the reference band of 22.5-25.0. For those with a BMI below 15.0, classified as severe thinness by the WHO, the hazard ratio rose to 2.16 (1.88-2.48) for men and 1.89 (1.62-2.20) for women. These are large, statistically robust associations.

Respiratory Deaths Show the Strongest Link

The risk was not uniform across causes of death. The strongest link was with death from respiratory illness. For the BMI 15.0-18.5 group, the hazard ratio for respiratory deaths was 2.16 (1.88-2.49) for men and 1.78 (1.52-2.09) for women. This is a higher relative risk than for deaths from all causes, and it is a finding that has been replicated in subsequent cohort analyses.

Bone Fractures

The epidemiological associations extend beyond death rates into specific illnesses. Bone health shows one of the most consistent findings. A 2005 meta-analysis by De Laet et al. in Osteoporosis International pooled 12 prospective cohorts with 59,644 men and women. It found that a BMI of 18.5 compared to 25.0 carried a relative risk of 1.95 (95% CI 1.71-2.22) for any fracture and 2.22 (95% CI 1.84-2.68) for hip fracture specifically. The mechanism is straightforward: lower body weight means lower mechanical loading on bone, which is a stimulus for bone mineral density maintenance. Underweight individuals tend to have lower bone mineral density. Less soft tissue padding around the hip means less force absorption in a fall.

Immune Function and Tuberculosis

Immune function also shows a clear gradient. A 2010 systematic review by Lönnroth et al. in the International Journal of Epidemiology covering 1,700,000 participants across six cohort studies found that a BMI below 18.5 was associated with a doubled risk of active tuberculosis (relative risk 2.0, 95% CI 1.5-2.6) compared to the reference range of 18.5-24.9. Malnutrition, even at subclinical levels, impairs cell-mediated immunity and reduces the body's ability to control mycobacterial replication.

Surgery and Recovery

Post-surgical outcomes are worse at low BMI. A 2009 analysis by Mullen et al. of 118,707 patients from the National Surgical Quality Improvement Program database found an odds ratio for death within 30 days of 1.52 (95% CI 1.33-1.74) for those with a BMI below 18.5 compared to the reference range of 18.5-24.9. Surgeons refer to this as the low-BMI risk in preoperative risk stratification, separate from the obesity-related risks.

Muscle Loss in Ageing

In older adults, a low BMI is a marker for sarcopenia, the age-related loss of muscle mass and function. The European Working Group on Sarcopenia (EWGSOP2, 2019) includes low muscle strength and low muscle quantity in its diagnostic criteria: grip strength below 27 kg in men and 16 kg in women, and appendicular skeletal muscle mass per height squared below 7.0 kg/m² in men and 5.5 kg/m² in women. A BMI below 18.5 in an older adult is often accompanied by cachexia, the wasting syndrome driven by chronic inflammation, which makes the weight loss both a cause and a consequence of declining health.

Outcome	Risk Estimate	Source
Death from any cause (men, BMI 15–18.5)	HR 1.47 (1.39–1.55)	Global BMI Mortality Collaboration, 2016
Death from any cause (women, BMI 15–18.5)	HR 1.33 (1.26–1.41)	Global BMI Mortality Collaboration, 2016
Death from respiratory illness (men, BMI 15–18.5)	HR 2.16 (1.88–2.49)	Global BMI Mortality Collaboration, 2016
Any fracture (BMI 18.5 vs 25.0)	RR 1.95 (1.71–2.22)	De Laet et al., 2005
Hip fracture (BMI 18.5 vs 25.0)	RR 2.22 (1.84–2.68)	De Laet et al., 2005
Tuberculosis (BMI < 18.5 vs 18.5–24.9)	RR 2.0 (1.5–2.6)	Lönnroth et al., 2010
Death within 30 days of surgery (BMI < 18.5)	OR 1.52 (1.33–1.74)	Mullen et al., 2009
Ovulatory infertility (BMI < 17.0)	RR 2.2 (1.6–3.1)	Rich-Edwards et al., 1994
Low birth weight infant (maternal BMI < 18.5)	OR 1.64 (1.38–1.94)	Han et al., 2011
Preterm birth (maternal BMI < 18.5)	OR 1.32 (1.20–1.45)	Han et al., 2011

Here is the single most important methodological point in the underweight literature, and the one most popular accounts get wrong. The association between low BMI and higher death rates is heavily confounded by reverse causation and smoking. Reverse causation means that pre-existing illness causes weight loss, not that low weight causes illness. A person with undiagnosed cancer, chronic obstructive pulmonary disease, or heart failure often loses weight before diagnosis. By the time they enter a cohort study, they have a low BMI and a high risk of dying. The weight loss is a symptom, not a cause.

Smoking is a second major confounder. Smokers tend to have lower body weight than non-smokers. Nicotine suppresses appetite and increases metabolic rate. Smokers also die more from cardiovascular illness and cancer. Compare low-BMI smokers to normal-BMI non-smokers and you are comparing two groups that differ in both weight and smoking status. The difference in deaths may be driven by the cigarettes, not the thinness.

The Global BMI Mortality Collaboration attempted to address this. When they restricted their analysis to participants who had never smoked and who did not have pre-existing illness or cancer, the elevated death rates at low BMI were substantially attenuated. The J-shape flattened at the low end. This does not mean the low-BMI risk disappears entirely. In some analyses a residual elevation remains. But it means a large fraction of the observed risk is artefactual. Studies that fail to adjust for these confounders overstate the danger of being underweight.

This is not a niche debate. The obesity paradox, the observation that higher BMI is sometimes associated with lower death rates in certain illness cohorts, is the same statistical problem viewed from the other side of the BMI distribution. Both are explained by the same confounders: reverse causation, smoking, and selection bias. The causal effect of adiposity on death, as estimated by Mendelian randomisation studies that use genetic variants to separate cause from consequence, is more linear than the observational J-curve implies. At the low end, the causal evidence for harm from low body fat is weaker than the observational evidence.

Why Bones Weaken

The biological pathways linking low body weight to poor health are worth describing because they clarify which effects are causal and which are confounded. Bone health is the most straightforward. Bone is mechanosensitive tissue. The primary stimulus for maintaining bone mineral density is the mechanical load placed on it by body weight and muscle contraction. A person with a low BMI has less load, less osteoblast activity, and lower bone mineral density. This is a direct causal chain. It is why fracture risk rises even in healthy thin individuals.

How Immunity Suffers

Immune function operates through a different route. Adipose tissue is not inert. It produces leptin, which regulates appetite and also modulates immune responses. Severe energy restriction and low fat mass reduce leptin levels, and with them the activity of natural killer cells and T-lymphocytes. The result is impaired immune surveillance. This increases susceptibility to infections, particularly intracellular pathogens like Mycobacterium tuberculosis. It is one reason why tuberculosis disproportionately affects underweight populations, and why nutritional rehabilitation is a standard part of tuberculosis treatment.

The Respiratory Connection

Deaths from respiratory illness fit both causal and confounded patterns. Low BMI is associated with reduced respiratory muscle strength and lower lung volumes, which can impair cough effectiveness and airway clearance. But chronic lung conditions like COPD themselves cause weight loss through increased metabolic demand and systemic inflammation. This creates a vicious cycle where the low BMI is both a consequence and an amplifier of the underlying illness. That is why the Global BMI Mortality Collaboration found the strongest low-BMI association for respiratory deaths, and why restricting the analysis to never-smokers significantly reduced that association.