Five WTC Health Program clinical centers of excellence serve cohort members in the New York–New Jersey–Pennsylvania metropolitan statistical area. Since July 1, 2002, the clinical centers conduct periodic health monitoring and provide diagnostic procedures and treatment for the GRC, a large, longitudinal, open (continuing to enroll eligible members) cohort. Since then, sociodemographic status and WTC response-related exposure information have been collected and reported by clinical staff using standardized questionnaires at members’ first health evaluation visit 1 (V1). Information on self-reported health status, diagnostic procedures, GERD treatment including medication use, and information from physical and laboratory examinations are ascertained at V1 and subsequent visits.

The WTC Health Program research was approved by the Institutional Review Boards of the GRC Data Center and Clinical Centers of Excellence. The research was conducted in concurrence with the principles of the World Medical Association Declaration of Helsinki (1975, revised 2013) and national and institutional committees’ standards regarding the conduct of human studies. Written informed consent was obtained for all members included in the study.

Of 52,641 GRC members who attended Program monitoring visits through December 31, 2022, 48,813 (92.7%) provided written, voluntary consent for data aggregation and research use. Of those consenting, this study excluded members who: never reported GERD (n = 24,407); reported GERD before September 11, 2001 (n = 1928); or did not report the earliest date they experienced GERD (n = 492). To assess the associations between diagnostic procedures, treatment, and GERD resolution, additional exclusions were 1758 members who attended only one monitoring visit, 773 with no longitudinal GERD data (e.g., beyond its first being reported), 379 with less than 1 year of GERD follow-up and nine who reported some GERD resolution but whose dates of resolution were recorded as occurring before the earliest dates of GERD diagnosis, producing a study sample of 19,067 GERD cohort members.

The study primary outcomes are resolution of self-reported GERD and self-reported use of health services for GERD including diagnostic procedures (endoscopy, laryngoscopy, nighttime pH monitoring), behavior modification (dietary modification or bed head-elevation), medication use (proton pump inhibitors [PPI], histamine 2 receptor antagonists [H2], and antacids) [9, 15–20, 26, 27].

This study analyzed data collected from interview questionnaires and physical examinations. Since 2007, these data have been directly entered into electronic databases by clinic staff.

A positive response to any of the following questions: “Have you ever been diagnosed with GERD,” “Since we last saw you, have you been diagnosed with GERD?,” “Do you still have GERD?,” and “Are you currently under the care of a doctor for GERD?” was defined as a diagnosis of GERD. GERD diagnostic procedures and their occurrence date were also self-reported. For members reporting GERD, the interviewer-assisted questionnaire asked “Since [September 11, 2001/your last exam have you made/do you still use] diet modifications/elevate head [in bed].” Details of types of dietary modifications and reasons for weight loss were not captured. Similarly, information on endoscopy and laryngoscopy was captured from questions “Since… endoscopy/laryngoscopy” and “Date of last” procedure. We do not report on hospitalizations or surgeries other than endoscopy or laryngoscopy as their purpose (to diagnose or to treat GERD) was not specified. Current use of medications was self-reported by brand or generic medication name at each visit. Except for “other antacids,” we assessed only GERD medications approved on the Program’s formulary per the January 2023 Program’s Pharmacy Benefits Manager.

Using the study’s longitudinal data, three types of GERD resolution were defined: resolved (without relapse)—when resolution was reported after its first diagnosis and GERD was not subsequently reported (Formula 1); intermittent GERD—when the earliest resolution was reported after first diagnosis and before a subsequent report of GERD; and unresolved—when no resolution was reported (Formula 2). Per the study inclusion criteria, only incident cases of GERD occurring on or after September 11, 2001, who had longitudinal data and GERD follow-up of more than 1 year were assessed. The analyses assessed GERD duration, which was calculated for each GERD resolution status as follows: no resolution—the interval between date of earliest report of GERD and last visit date; intermittent GERD—the interval between date of earliest report and latest report of GERD; and resolved GERD—the interval between date of earliest report of GERD and date of resolution. GERD follow-up was calculated as the interval between earliest date of reported GERD and last monitoring visit date.

Formula 1: Resolved = All reported GERD dates < first resolution date

Formula 2: Intermittent = First resolution date (>first reported GERD date & <last reported GERD date)

The temporality of “exposure” to a GERD diagnostic procedure or therapy was assigned according to a member’s resolution status. Members with no resolution that reported a procedure or therapy were classified as “exposed” to the procedure or therapy, both during and after their unresolved GERD. Members with any GERD resolution that reported diagnostic procedures or therapies before or on the earliest date of GERD resolution were classified as having been used during GERD (e.g., before resolution), while diagnostic procedures or therapies reported after the earliest date of GERD resolution were classified as “exposed after GERD resolution.” Members who did not report a procedure or therapy were classified as unexposed during and after GERD resolution.

Descriptive statistics were generated to characterize the sample, diagnostic procedures as reported from the second monitoring visit (when this information was first captured) onwards, treatments and Program-certified GERD for members who completed monitoring visits since July 1, 2011, when GERD became a Program-certified condition, and the dependent variable, resolution of self-reported GERD. Cross-tabulations of sociodemographic characteristics and GERD treatments reported are presented with two-tailed chi-square 95% probability values. Independent binomial logistic regressions were used to assess the associations of treatment with intermittent GERD and resolved GERD, compared with members reporting no resolution. Four regression models were assessed. Model 1 adjusted for the major potential confounders for GERD including age, sex, race (black, Asian, other, not reported, compared to the referent category, white), Hispanic ethnicity (compared to non-Hispanic or those not reporting ethnicity), measured body mass index (BMI 25–29 considered as overweight, ≥ 30 considered as obese, compared to < 25) calculated as 703× (pounds/inches²]), and marital status (married/partnered, other, compared to single) at visit 1, occupation (construction, electrical/telecommunications, and other installation/repair, transportation and material moving, other jobs, compared to protective services) on September 10, 2001, lifetime cigarette smoking status (former, current, compared to nonsmoker), and WTC exposure on September 11, 2001 when toxic exposures were theoretically greatest (compared to afterwards) [3]. Model 2 included model 1 variables, Program GERD certification, GERD follow-up duration, and pre-resolution endoscopy or laryngoscopy, bed head-elevation, dietary modification, and use of PPI, H2, or Program-approved antacids. Model 3 included model 2 variables, Barrett’s disease in GERD-certified members and Program certification for the most severe consequence of GERD, esophageal cancer. Model 4 included model 3 variables and two-way interaction terms for age, GERD diagnostic procedures, dietary modifications and use of PPI, H2, and Program-approved antacids. Sensitivity regression analyses were conducted, limited to GERD certified members. The regression analyses are presented with adjusted odds ratios (AOR) and two-tailed 95% confidence intervals (CI). Hosmer–Lemeshow (HL) goodness-of-fit statistics were evaluated [28]. Figure 1 illustrates the use of GERD treatments during and after GERD resolution in members with intermittent and resolved GERD. The analyses were conducted using SPSS for Windows version 28.0.1.1(15) (IBM Corp., Armonk, NY).

There were 19,067 consenting, study-eligible members with self-reported incident GERD, 5.8% of whom had Barrett’s disease and 0.1% esophageal cancer. Most were married or partnered, white, men, nonsmokers, overweight or obese at their first monitoring visit, and between 32 and 44 years old when they reported GERD first occurred (Table 1). Ninety-seven percent attended Program monitoring visits on or after July 1, 2011 when certification of members for WTC-related conditions began. Forty-four percent of the members with GERD were onsite responders on September 11, 2001. Fifty-nine percent were certified for GERD (considered WTC related-GERD). Approximately half of the members had diagnostic endoscopy; almost none had diagnostic laryngoscopy or nighttime pH monitoring. Proton pump inhibitors and dietary modification were the most commonly used treatments, although many people reported they elevated their heads while in bed and took Program-approved antacids. While some differences between the entire sample (GRC members with self-reported GERD) and the subgroups with visits on or after July 1, 2011 and GERD certified members were statistically significant due to large sample size, most were not substantively different (Table S1). Attending a monitoring visit on or after July 1, 2011 had little substantive influence on the members’ use of health services for GERD (Table S1). Members who were certified for GERD had higher rates of Barrett’s disease, diagnostic endoscopy, bed head-elevation, and dietary modification, and slightly higher rates of proton pump inhibitor use than did all study members (Table S1).

Most (81.8%) did not report any GERD resolution, but 12.6% reported intermittent GERD resolution and 5.5% reported resolution without relapse (Table 1). Most members reporting diagnostic procedures or treatments for GERD experienced less GERD resolution than those not reporting these services or therapy. However, members who identified their race as black (OR = 1.53, [1.34, 1.73]), Hispanic (OR = 1.18, [1.08, 1.29]), and members with body mass index < 25 at their first visit (OR = 1.26 [1.13, 1.41]), experienced intermittent or resolved GERD significantly more often than those not reporting any resolution (unadjusted regression data not shown).

The average (mean ± standard deviation) GERD duration was 10.5 ± 6.0 years in members with no resolution, 12.0 ± 5.2 years in members with intermittent GERD and 5.6 ± 4.8 years in members with resolved GERD. Mean GERD follow-up was (identical to GERD duration at) 10.5 ± 6.0 years in members with no resolution, 13.5 ± 4.7 years in members with intermittent GERD and 10.6 ± 5.4 years in members with resolved GERD. Members who reported intermittent GERD increased their use of bed head-elevation, dietary modifications, H2s, and Program-approved antacids after first reporting some resolution, but decreased their use of PPIs after their earliest reported resolution (Figure 1). Conversely, those with resolved GERD reported less use of these therapies after GERD resolution (Figure 1).

The four adjusted logistic regression models produced similar results. Consequently, only results for model 3 are presented here unless otherwise specified. Greater age at GERD diagnosis was associated with more resolution (Table 2). Black and Hispanic members had greater odds of experiencing intermittent resolution and resolved GERD than white and non-Hispanic members, respectively. Compared with members with first visit BMIs <25, members with BMI ≥ 25, particularly those with BMI ≥ 30, had less GERD resolution. Members who implemented dietary modifications had significantly greater odds of intermittent resolution. However, members who were current smokers, Program-certified for GERD, had surgical diagnosis (endoscopy or laryngoscopy), used bed head-elevation or used PPIs or H2s for GERD reported less GERD resolution than their referent groups (respectively, nonsmokers, not GERD certified, had no surgical diagnosis, did not use bed head-elevation or GERD medications). The regression results were nearly identical for the indices of diagnosis and therapy in models with and (data not shown) without adjustment for patients with Barrett’s disease and esophageal cancer certification and for analyses limited to GERD-certified members.

Similarly, adjusted analyses including two-way interaction terms for GERD services and medications used with age at GERD diagnosis, race, Hispanic ethnicity, and for first visit BMI found people diagnosed ≥ age 40 who had endoscopy had less resolution (AORs 0.47–0.53, 95% CI 0.3–0.8, all p < 0.01) and people who implemented dietary modifications who also used PPIs (AOR 1.9, 95% CI 1.5, 2.5, p ≤ 0.001) or Program-approved antacids (AOR 1.7, 95% CI 1.2, 2.2, p ≤ 0.001) had significantly more GERD resolution (remaining interaction statistics were not significant and are not shown).

The Hosmer–Lemeshow goodness-of-fit statistic was significant for intermittent resolution but not (p = 0.21) for the smaller sensitivity analysis limited to GERD certified members, even though the outcome AORs were nearly identical (data not shown). The HL statistics were not significant for the smaller sample of resolved GERD (p = 0.14) or its sensitivity analysis in GERD certified members (p = 0.08).