Data

For this retrospective cross-sectional study, we analyzed hospital discharge data from Mississippi for 2016–2022. The Mississippi State Department of Health, in collaboration with the Mississippi Hospital Association, collects those data and shares them with the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project as part of the organization’s national hospital care database (5). The rich population-level data contain demographic, clinical, and financial information derived from billing claims submitted by all nonfederal acute-care hospitals in Mississippi.

Case Definition

For this study, we included infants ≤1 year of age living in Mississippi. To meet the goals of our investigation, we created 2 study groups. The first comprised all infants hospitalized with CS, including those diagnosed at delivery hospitalization and those who were admitted after delivery (postdelivery subcohort). We used data from this more inclusive group to study prevalence, trends, and demographics among infants hospitalized with CS. The second group (delivery subcohort) included only infants either delivered in the hospital or delivered outside of the hospital but admitted with a delivery code from the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM). We needed data from the delivery subcohort because we were interested in pregnancy outcomes associated with syphilis, including preterm and low-birthweight deliveries. To identify the delivery subcohort, we developed an algorithm that incorporated infant age and ICD-10-CM code Z38, the neonatal diagnostic code for birth. In hospital discharge data, neonatal patient age was based on date of admission. In our delivery subcohort, we included all infants whose ages were recorded as 0–3 days with day 0 being the day of birth.

Data Variables

We used ICD-10-CM code A50 to identify diagnoses of infant hospitalizations with CS. We referenced related ICD-10-CM codes to study newborns affected by maternal use of illicit substances: opiates (P04.14), amphetamines (P04.16), cocaine (P04.41), hallucinogens (P04.42), cannabis (P04.81), unspecified drugs of addiction (P04.40, P04.49), and neonatal abstinence syndrome (P9.61; i.e., severe withdrawal symptoms). For preterm deliveries, we referenced codes P0.72 for extreme newborn immaturity (<27 completed weeks of gestation) and P0.73 for premature births (<36 completed weeks of gestation). We referenced ICD-10-CM code P22 to screen the data for newborn respiratory distress, which were of special interest for this study because this condition is a major cause of neonatal illness and death (6). To evaluate birthweight, we used the variable for birthweight in the dataset. Likewise, we detected deaths of infants with diagnosed CS based on the dataset variable for discharge status, which includes in-hospital deaths. Finally, we used neonatal ICD-10-CM code P95 to identify stillbirths and screened those records for CS diagnoses.

For our research, we collected key demographic characteristics, such as urbanicity, race, and socioeconomic status, for each individual CS patient. We defined urban versus nonurban status on the basis of the CDC National Centers for Health Statistics urban–rural classification scheme (7), in which counties with >50,000 inhabitants are classified urban and counties with <50,000 nonurban (10,000–49,999, micropolitan; <10,000, rural). We categorized data on race into 3 groups: African American, White, and others (because few infants with CS in the study were not either African American or White). The following racial categories are included in the hospital discharge dataset: White, African American/Black, American Indian/Alaska Native, Asian, Native Hawaiian/other Pacific Islander, multiracial, declined, and race unavailable/unknown. We were unable to examine ethnicity because few infants of Hispanic origin were included in the CS cohort. We used Medicaid insurance coverage as a proxy for low socioeconomic status.

Statistical Analyses

We used data from individual patients as units of analysis. To select individual patients, we used unique identifying numbers and flagged the first hospitalization for each patient. We included only data from first admissions, not transfers or readmissions. Our analyses included descriptive and inferential statistics. We compared demographic and clinical characteristics between patients with and without CS using independent t-tests for continuous variables and χ² tests for categorical variables. To evaluate the strength of association between different demographic and clinical factors and CS, we performed multivariable regression and obtained adjusted odds ratios (aORs) (8). We conducted all analyses using SAS 9.2 statistical software (9).

Trend Analyses

The number of infants hospitalized with CS spiked by 1,000% over the study period, rising sharply from 10 infants in 2016 to 110 in 2022 (Figure 1, panel A). The rate of hospitalization of infants with CS increased from 2.0/10,000 total infant hospitalizations in 2016 to 24.8/10,000 total infant hospitalizations in 2022, representing a 1,140% increase (Figure 1, panel B). Additional analyses revealed 2 notable trends. First, the rate increase in the number of infants hospitalized with CS varied by race. Although infants with CS were overwhelmingly African American, the magnitude of increase was higher among White infants. Among African American infants, CS hospitalizations increased by 1,029%, rising from 7 infants in 2016 to 79 in 2022. Over the same time period, the admissions among White infants jumped by 2,600%, from 1 infant in 2016 to 27 in 2022. Second, hospitalization for infants with CS started to increase in 2018, before the onset of the COVID-19 crisis; the pandemic, however, accelerated this uptrend.

Demographic Risk Factors

Bivariate analyses of associations between CS and different demographic characteristics revealed that infants with CS were more likely than those without to be African American (71.1% vs. 43.3%; p<0.001), covered by Medicaid (92.6% vs. 63.9%; p<0.001), and >3 days of age (7.4% vs. 2.0%; p<0.001). After adjusting for covariates, Medicaid insurance coverage (aOR, 5.24, 95% CI, 3.58–8.00) was the strongest independent demographic risk factor; however, infants >3 days of age (aOR, 3.82, 95% CI, 2.52–5.55) and African American race (aOR, 2.26, 95% CI, 1.37–4.06) also retained associations with congenital syphilis risk (Table 1).

Clinical Characteristics of the Delivery Cohort

We compared findings on clinical characteristics from this analysis between infants with and without CS (Table 2). In the bivariate analysis, newborns with CS were significantly more likely than those without CS (21.5% vs. 2.4%; p<0.001) to have a maternal illicit drug use ICD-10-CM code on record. Among newborns with CS, 73 were affected by maternal substance use; however, only 36/73 (49.3%) records for those newborns included a substance-specific diagnostic code. Among those 36 infants, maternal use of cannabis was recorded for 69.4%, cocaine 27.8%, amphetamines 25.0%, and opioids 2.8%. Drug-use categories were not mutually exclusive, and 7/36 CS-infected newborns with recorded substance-specific diagnostic codes were exposed to >1 substance in utero. Neonatal abstinence syndrome was present in 8 cases.

In the multivariable analysis of clinical characteristics, the variable with the strongest association with CS was maternal substance use. Even after controlling for demographic risk factors, the CS cohort had almost 10 times higher odds of being affected by maternal illicit drug use than the cohort without CS (aOR, 9.39, 95% CI 7.16–12.16). Aside from minor variations, the proportion of maternal illicit drug use among the CS cohort remained high throughout the study period. In 2016, the year with the highest proportion of substance-exposed infants, 25% of all newborns with CS also had a diagnostic code indicating maternal use of illicit drugs. During the last 3 years of the study period, there was a gradual but steady increase in the proportion of newborns hospitalized with CS and exposed to illicit substances in utero: 17.7% in 2020, 22.7% in 2021, and 23.8% in 2022 (Figure 2).

Infants in the CS cohort were also more likely to suffer from severe complications, including prematurity and low birthweight, than infants in the cohort without CS. Among newborns with CS, 92 (27.1%) were premature, compared with 43 (12.7%) for newborns without CS. Among premature infants with CS, 8/92 were extremely premature (i.e., born before completing 28 weeks of gestation). Newborns with CS had >2 times higher odds of being born prematurely than did newborns without CS (aOR 2.26, 95% CI 1.77–2.86) after controlling for demographic risk factors. Among 87 newborns with CS who had birthweights <2,500 g, 59 had low birthweight (LBW) of 1,500–2,499 g and 28 had very low birthweight (VLBW) of <1,500 g. Compared with the cohort without CS, infants in the CS cohort were more likely to be LBW (17.9% vs. 9.7%; p<0.001) and VLBW (8.5% vs. 1.9%; p<0.001). Alarmingly, odds for newborns with CS to be LBW were nearly 2 times higher (aOR 1.81, 95% CI 1.35–2.40) and to be VLBW >4 times higher (aOR 4.05, 95% CI 2.67–5.90) than for newborns without CS. On average, newborns with CS weighed 349 g less (mean birthweight 2,788 g) than newborns without CS (3,137 g; p<0.001) (Figure 3). Newborn respiratory distress was also more common among the CS cohort (21.2%) than for the cohort without CS (9.1%; p<0.001). Given the well-established connection between prematurity and respiratory distress in newborns, it is unsurprising that 1 in 5 premature newborns with CS in our study was also diagnosed with respiratory distress. Among the 72 newborns with diagnoses for CS and respiratory distress, 48 (66.7%) also had a diagnostic code for preterm delivery.

Use of Resources

In terms of resource use, infants with CS had a longer mean hospital stay (14.5 days vs. 3.8 days; p<0.001) and mean hospital charges ($56,802 vs. $13,945; p<0.001) than did infants without CS. During 2016–2022, hospital charges for the CS cohort totaled $20,846,196; Medicaid-insured infants accounted for $19,444,608 (93.3%) of those charges.

Mortality

During the last 4 years of the study period (2019–2022), 6 infants with CS died in hospital, half of those during the last year of the study period (2022). Among the infants with CS who died, 83.3% had extremely low birthweight (<1,500 g), and 66.7% were extremely premature. Among our study population, no records indicated both stillbirth and CS.