Obesity is a major and growing public health problem in the United States, affecting more than 40% of adults and roughly 1 in 5 children and adolescents. Early-onset and prolonged obesity carry increased long-term morbidity and mortality. While infections are established drivers of some chronic conditions and classically linked to long-term weight loss (for example, tuberculosis and parasitic disease), several studies have reported a less intuitive association between early-life infections and later increased body mass index (BMI) or obesity. Previous work has also raised the possibility that antimicrobial drug exposure in infancy might influence later weight, but disentangling antimicrobial effects from the underlying infection is challenging. This investigation within the Kaiser Permanente Southern California (KPSCAL) system sought to confirm prior observations, quantify absolute weight differences commensurate with obesity rate differences, and assess the role of antimicrobial treatment.
The study identified a retrospective, longitudinal cohort of infants born 2008–2012 within KPSCAL. Exclusion criteria included birthweight <2,500 g, gestational age <37 weeks, lack of KPSCAL membership during the first 2 years of life, absence of primary care visits in each of the first 2 years, missing outcome measurements at the 4–6 and 10–15 year time points, nonbinary gender identity, and instances of antimicrobial drug exposure without a concurrent infection diagnosis.
KPSCAL provides integrated inpatient and outpatient care across Southern California; the authors note that most member care occurs within the system, although care obtained outside the system (for example, while traveling or for certain specialized procedures) might not be captured in the electronic medical record.
Investigators classified infants into three mutually exclusive groups:
Infections were considered distinct if recorded more than 2 weeks apart. Multiple infection diagnoses occurring within a single 2-week window were treated as one clinical infection. Antimicrobial drug exposure was further categorized as narrow-spectrum (amoxicillin, penicillin, first-generation cephalosporin, or combinations of these) or broad-spectrum (any other antimicrobial agent or combinations).
Timing of first infection (<6 months vs 6–12 months), number of infections, timing of first antimicrobial exposure (<6 vs 6–12 months), and number of antimicrobial courses were recorded. The IWAT group captured the number of infections separately from the number of antimicrobial drug courses, enabling exploration of frequency effects.
Primary outcomes were obesity status (BMI >95th percentile) and absolute weight (kg) measured at two discrete childhood intervals: ages 4–6 years and 10–15 years. For analysis, investigators used the BMI and weight measurements closest to age 5 for the 4–6 year group and closest to age 10 for the 10–15 year group.
BMI values were standardized by age and sex using CDC growth-chart tools. The authors deliberately emphasized absolute weight rather than z-scores because absolute weight differences are more directly interpretable and clinically meaningful; sex and height were included as covariates in multivariable analyses to account for body composition differences.
The article describes use of multivariable regression to assess associations between infant infection exposure groups and later weight outcomes, accounting for covariates including sex, height, and age group. The analysis examined dose–response relationships by number of medically attended infections and evaluated the predictive contribution of antimicrobial treatment by comparing IWOT and IWAT groups. The provided text does not include full details of statistical models, specific covariates included beyond sex and height, model diagnostics, or effect estimates (see Limitations below).
Multivariable regression reportedly showed that infants with early medically attended infections had higher rates of obesity at ages 4–6 years and 10–15 years compared with infants without such infections.
The association was proportional to the number of infections diagnosed, and was accompanied by statistically significant increases in absolute weight.
Most infections in the cohort were viral in nature.
Compared with infants whose infections were not treated with antimicrobial drugs, antimicrobial treatment was less predictive of later weight outcomes.
The investigators concluded there was a correlation between infections diagnosed in infancy and later weight outcomes but stated they could not determine causal mechanisms from the data presented.
Note: the provided source text includes a detailed Methods section and an abstract summary of results, but does not include the full results tables, numerical effect sizes, confidence intervals, or p-values in the excerpt available here.
The manuscript situates its findings within prior literature: a prospective study in the Philippines reported an association between early upper respiratory infections and increased adult BMI, and several retrospective studies in the United States have linked early antimicrobial exposure with later obesity. Other work has suggested that antimicrobial exposure may serve as a proxy for underlying infection rather than being the causal factor for later weight gain. The present study sought to separate the contribution of infection itself from antimicrobial treatment by comparing infected infants with and without antimicrobial exposure.
The authors also discuss methodological concerns raised previously about using obesity as a binary outcome (BMI threshold) and note the value of reporting absolute weight increases as a complementary, clinically interpretable metric.
The provided article excerpt does not contain full numeric results (effect sizes, confidence intervals, or p-values), the composition of covariates in regression models, subgroup analyses, or results stratified by infection type beyond the statement that most infections were viral. Those details are not reported in the source text excerpt supplied here.
As a retrospective EMR-based study, unmeasured confounding (socioeconomic factors, diet, physical activity, home environment, parental BMI, breastfeeding, and other early-life exposures) may influence associations; the excerpt does not specify which of these were measured or adjusted for.
The authors explicitly state they could not determine causation from the observed correlations.
The KPSCAL system captures most but not necessarily all care for members; infections or antimicrobial exposures occurring outside the system could be missed.
Within this retrospective KPSCAL birth cohort (2008–2012), medically attended infections in infancy were correlated with increased absolute weight and higher obesity rates at ages 4–6 and 10–15 years, with a dose–response relation to the number of infections. Most infections were viral, and antimicrobial treatment was less predictive of later weight outcomes than the presence and number of infections. The study corroborates prior observations that early clinically attended infections are associated with later weight gain but does not establish causality. Full numerical results and certain analytic details were not available in the provided text.
If you would like, I can obtain and summarize specific numerical findings, effect estimates, and tables/figures from the full published article if you provide the remainder of the text or permit me to access the full paper.