How do you choose a safe infant formula, and are homemade formulas okay?

What "regulated" means. In the US, the FDA requires minimum amounts of 30 nutrients in routine infant formula and maximums for 10 that can harm in excess, such as vitamins A and D. Makers must register, notify the FDA before changing a formula, and test every batch for Cronobacter². Specialized formulas for medical conditions are exempt from some of these rules, so use them only on a clinician's advice². Outside the US, check that a product meets your national standard.

Types. Routine formulas can be cow's-milk, goat's-milk or soy based. Partially hydrolyzed, thickened "spit-up" and lactose-free versions exist for common digestive complaints, while extensively hydrolyzed, amino-acid, premature and metabolic formulas are for medical needs. The American Academy of Pediatrics recommends formula with iron to prevent iron-deficiency anemia¹. Ask your pediatrician before switching types.

What the extras do.

• Human milk oligosaccharides (HMOs). One trial of 175 infants found HMO formula supported normal growth and fewer parent-reported chest infections, which still needs confirming⁸.
• DHA and ARA oils. Trials have not shown consistent developmental benefits for term babies⁹.
• Carbohydrates. Lactose is the main sugar in human milk¹. Formula with corn syrup solids was linked to modestly higher obesity risk in one large observational study¹⁰.
• Goat milk. Four industry-funded trials found growth and safety similar to cow's-milk formula¹¹.
• Organic. Direct comparisons are scarce, so health advantages remain unproven¹².
• Contaminants. Most of 312 samples FDA tested had undetectable or very low levels of toxic elements, pesticides, PFAS and phthalates⁶.

Preparing it safely. Cronobacter, a rare but serious cause of infant meningitis and sepsis, can get into powdered formula after the container is opened. Babies two months or younger, born prematurely or with weakened immunity are most at risk, and sterile liquid formula is safer for them^2,7. Wash your hands, sanitize bottles, use safe water, and follow the label for mixing, storage and discarding leftovers².

Homemade and raw-milk formulas. The FDA and the American Academy of Pediatrics advise against both^1,3. Three infants fed homemade "alkaline diet" formulas developed severe low calcium and rickets⁴.

Age limits. Cow's milk is not a substitute for formula in the first year, and "toddler" drinks are not for babies under 12 months^1,5.

Regulation and contaminant testing

FDA's 2023 to 2025 market-basket survey tested 312 samples from 16 brands (278 powdered, 11 concentrated liquid, 23 ready-to-feed; 258 cow's-milk, 44 soy, 10 amino-acid based). Mercury was undetected in 95% of samples. Lead ranged up to 1.1 ppb, with 95% of samples under 0.5 ppb; cadmium reached 1.5 ppb; arsenic, undetected in only 6%, reached 4.7 ppb (95% under 2.0 ppb) and was inorganic in the subset speciated. Of 318 pesticides tested, 99% of samples had none detectable and glyphosate was never found; 25 of 30 PFAS were never found; phthalates were undetected in 54%. FDA says detection does not make a product unsafe and calls US formula safe, while noting that snapshots miss lot-to-lot variation and some samples triggered follow-up⁶.

HMOs, oils and sugars

Puccio randomized 175 healthy newborns to a cow's-milk formula with or without 2'-fucosyllactose (1.0 g/L) and lacto-N-neotetraose (0.5 g/L) for six months. Weight gain was non-inferior (difference -0.30 g/day, 95% CI -1.94 to 1.34). Parent-reported bronchitis was lower through 12 months (10.2% versus 27.6%) and antibiotic use lower (42.0% versus 60.9%), but these were secondary outcomes, and the authors called for confirmation⁸.

Simmer's Cochrane review of 15 trials (1,889 term infants) found most showed no neurodevelopmental benefit from added DHA and ARA, visual acuity benefits in four of nine trials, and possibly marginally lower weight at one year. It could not recommend routine supplementation⁹.

Anderson analyzed 15,246 WIC children given cow's-milk formula. Those issued corn-syrup-solids formula had 10% higher obesity risk at age 2 (RR 1.10, 95% CI 1.02-1.20), 16% higher after 12 months of exposure (RR 1.16, 95% CI 1.05-1.28), and 7% higher at age 4. The data are administrative and observational; formula choice may track a baby's digestive symptoms or family circumstances¹⁰.

Goat milk and organic

Jankiewicz pooled four randomized trials (670 infants; all industry-funded, each with some risk-of-bias concern). Goat-milk formula matched cow's-milk formula for weight (z-score difference 0.21, 95% CI -0.16 to 0.58), length and head circumference, with similar stool frequency and adverse events¹¹. For organic formula, Al-Beltagi found comparative studies scarce and formula differences a major limit; suggested advantages in lipid profile, methionine content and antibiotic-resistant bacteria are potential, not demonstrated health gains¹².

Cronobacter

Strysko reviewed 183 invasive infant infections (1961 to 2018). Two-thirds were in neonates, 38% (42 of 112) died, and 79% (81 of 102) had recently consumed powdered formula. Cases rose after 2004 (about 8.7 a year worldwide), and among US cases 56% were full-term and 78% non-hospitalized. Contaminated formula, most often from an opened container, was found in 30% (21 of 71) of investigations⁷. These infections are rare but severe, which is why preparation rules and sterile liquid formula for the highest-risk babies matter.

Homemade and raw-milk formulas

The AAP statement notes that raw milk from cows, goats and sheep carries Listeria, Campylobacter, Salmonella, Brucella and E. coli O157; that the claimed benefits of raw milk have not been shown in evidence-based studies; and that pasteurized milk gives equivalent benefits without the infection risk³. Calello's MMWR report describes three infants aged 4, 5 and 9 months fed homemade "alkaline diet" formulas of sea moss, hemp seeds, coconut water or milk and dates. Two had calcium of 4.0 and 4.5 mg/dL (normal 8.8 to 10.8), all had rickets, one had a cardiac arrest with hypoxic brain injury, and one had severe iodine deficiency⁴. Case reports cannot show frequency, but the missing nutrients are ones infants cannot do without, and FDA reports hospitalizations from homemade formula¹.

Cow's milk before 12 months

Griebler's meta-analysis of 23 studies found infants given cow's milk had 3.76 times the risk of iron-deficiency anemia compared with follow-on formula (95% CI 2.73-5.19; four studies)⁵.

Where the evidence is thin

Every ingredient claim rests on one trial, a mixed Cochrane review, observational data or industry-funded studies. Nobody can ethically randomize babies to homemade formula, so that advice rests on regulation, case reports and biology.

Bottom line (confidence 3 of 5)

Regulated formula is safe, preparation rules are well founded, and homemade or raw-milk formula is dangerous. Differences between regulated brands are small and poorly proven, so choose a type with your pediatrician and prepare it exactly as labelled.