Obesity Surgery

, Volume 20, Issue 7, pp 861–870

The Impact of Bariatric Surgery on Comorbidities and Medication Use Among Obese Patients


    • Analysis Group
  • Severine Ledoux
    • Service des Explorations Fonctionelles—Hôpital Louis Mourier (AP-HP), Colombes and UFR de MédecineUniversité Paris 7
  • Christine Clerici
    • Service des Explorations Fonctionelles—Hôpital Louis Mourier (AP-HP), Colombes and UFR de MédecineUniversité Paris 7
  • Francois Cremieux
    • Hôpital Hotel-Dieu (AP-HP)
  • Marric Buessing
    • Analysis Group
Clinical Research

DOI: 10.1007/s11695-010-0163-6

Cite this article as:
Crémieux, P., Ledoux, S., Clerici, C. et al. OBES SURG (2010) 20: 861. doi:10.1007/s11695-010-0163-6



The risks and benefits of bariatric surgery have rarely been evaluated in large multiyear patient samples. This study identifies the short- and long-term impact of bariatric surgery on comorbidities and medication use among obese patients.


A comprehensive analysis of 5,502 obese patients who underwent bariatric surgery was performed. Submissions of reimbursement claims, including diagnostics and medication use, were compared in the 90 days preceding the surgery and 30 up to 1,110 days following the surgery. Presurgery and postsurgery frequency counts were performed on diagnostics and medication use to identify trends.


Among 5,502 patients, significant decreases in the prevalence of reported comorbidities were observed during the short-term postsurgery period and sustained for up to 3 years of follow-up. Compared to the presurgery period, significant decreases (p < 0.05) were observed after 3 years for total cardiovascular disorders (43.6% vs. 14.2%), diabetes mellitus (19.9% vs. 7.7%), chronic obstructive pulmonary disease and other respiratory conditions (57.7% vs. 16.2%), diseases of the musculoskeletal system and connective tissue (32.6% vs. 27.7%), and mental disorders (30.7% vs. 14.8%). Over the same period, the frequency of medication use decreased significantly for a number of conditions including infections, pain, respiratory, cardiovascular, gastroenterologic, lipidemic, and diabetic conditions. Anemia, however, increased from 3.8% to 9.9%, and use of nutritional supplements increased significantly.


Bariatric surgery was associated with significant reductions in reported claims for short- and long-term health outcomes and reduced medication use for major disease categories.


Morbid obesityQuality of lifeBariatric surgeryComorbidityCardiovascular diseaseDiabetes mellitus


With diet, exercise, and medication use relatively ineffective in maintaining long-term weight loss [1], bariatric surgery can induce long-term weight loss and may be considered for adults with a body mass index (BMI) ≥ 40 or a BMI ≥ 35 with serious coexisting conditions [2]. The American Society for Metabolic and Bariatric Surgery estimated that the number of Americans undergoing bariatric surgery rose from just over 20,000 procedures in 1997 to 220,000 in 2008 [3, 4].

Several case studies and meta-analyses have demonstrated improved comorbidities following bariatric surgery. For example, a meta-analysis of 22,094 patients revealed that bariatric surgery was associated with a mean loss of excess weight of 61.2% and marked improvements in several specific comorbidities, including diabetes, hypertension, obstructive sleep apnea, and hyperlipidemia [5]. However, very few studies have been conducted in large cohorts to evaluate overall health status or changes in medical use after bariatric surgery. An observational cohort study based on submitted claims for the reimbursement of services or medication showed that, after a mean of 5.3 years, bariatric surgery was associated with a mean excess weight loss of 67.1% [6]. The bariatric surgery cohort experienced significant relative reductions in the risk of cardiovascular, endocrine and respiratory diseases, cancers, infectious diseases, and psychiatric disorders. Conversely, the incidence of gastroenterological disorders was increased in surgical patients. Strikingly, bariatric surgery was associated with an 89% reduction in the relative risk of mortality. A comparison of the prevalence of specific comorbidities before and after vertical banded gastroplasty (VBG) surgery was performed in the Swedish Obese Study (“SOS”) Subjects Study. This study followed 641 obese patients who underwent surgery and matched controls over a 10-year period [7]. After 10 years, the group that received the VBG surgery, which is rarely performed in the US, maintained a 16.1% weight loss vs. a 1.6% weight gain in the control group. Surgery was associated with significant improvements in common comorbidities, including diabetes and dyslipidemia.

We are not aware of a long-term study which compares the use of medication among obese patients before and after surgery, thereby using patients as their own controls. Two recent study report reductions in short-term medication use using a pre- and post-bariatric-surgery comparison at 6 months and 1 year, respectively [8, 9]. Additionally, the SOS study reported a 77% higher use and cost of medications among obese patients compared to the general population [10].

Our study relies on a dataset of over 5,000 patients who underwent bariatric surgery to examine changes in reported comorbidities and medication use. We examine the evolution of reported comorbidities over a period of up to 3 years by comparing the prevalence of the most frequent conditions associated with obesity before and 30 to 1,110 days after surgery. Unlike previous studies which relied on patient-reported comorbidities often from single centers [10, 11], this analysis relies on claim-associated diagnostics from hundreds of institutions across the country and reflects actual clinical practice rather than procedures rarely performed in the USA (e.g., VBG) or observed under clinical trial conditions [12].

Materials and Methods

Data Collection

We used de-identified health insurance and disability claims for the reimbursement of services and medication from an Ingenix database of approximately five million employees, spouses, and dependents from 31 large companies throughout the USA. Collectively, these companies have operations nationwide in a broad array of industries and occupations, including financial services, transportation, technology, manufacturing, telecommunications, energy, and food and beverage. The database includes claims for inpatient and outpatient medical services (including diagnoses and procedures) and outpatient prescription medication dispensing records. Laboratory values, BMI, weight, and height were not available. However, claims associated with bariatric surgery as well as International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes identifying obesity and/or morbid obesity were clearly identifiable. The claims database covered the period from January 1, 1998, through January 31, 2006. Patient data included age, sex, type of health plan, and geographic region (northeast, south, west, Midwest).

Study Population Selection

Diagnosis-specific claims were identified by ICD-9-CM codes and prescription medication claims by National Drug Codes (NDC). Medical procedures were coded using Current Procedural Terminology (CPT) codes, Healthcare Common Procedure Coding System, and ICD-9-CM procedure codes.

Claims for patients undergoing bariatric procedures (case group) were identified by ICD-9-CM codes according to the following criteria: ≥1 CPT code for bariatric surgery (43644, 43645, 43845, 43846, 43847, 43842, 43843, S2085, S2082, S2083) and continuous insurance coverage and eligible for health care benefits for at least 3 months before and at least 4 months after the surgery. All patients with a code for bariatric surgery also had a diagnosis of obesity and other hyperalimentation (ICD-9-CM 278.0x).


For patients identified using the criteria defined above, the presurgery period was specified as starting with claims made up to 90 days before the date of the first observed surgery claim. The postsurgery period was defined by claims made 30 to 1,110 days after the first surgery claim. The postsurgery period was initiated 30 days after the surgery to eliminate biases resulting from immediate postsurgery medication use or early surgery-related complications.

Statistical Analysis

The main outcome variables were the number of patients with at least one claim for inpatient/outpatient services and prescription medication. The number of patients with ≥1 claim for each ICD-9-CM code and for each prescription drug in the presurgery period was identified. Claims describing the 50 most prevalent diagnoses and 50 most prescribed medications were used for subsequent analyses. Patients with ≥1 claim for each of the top 50 diagnoses and medications were identified before and after surgery. For each diagnosis, the percentage of patients with at least one claim in the 90 days preceding the surgery was compared with the percentage of patients with at least one claim in each 90-day postsurgery period, up to 1,110 days using chi-squared tests. Because the claims database covered 1998 through 2006, patients whose surgery occurred in 2003 or later were not observed for the full 3-year period. The comorbidity prevalence for the 1,372 patients who were observed for the full 3-year postsurgery period was compared with the comorbidity prevalence for the entire postsurgery sample to ensure that no bias was introduced by the timing of the surgery.

Medication use was identified using NDC and further classified using American Hospital Formulary Service (AHFS) therapeutic classes. A similar statistical analysis was performed on medication use. All statistical analyses were performed using SAS V9.1.


A total of 5,502 patients were identified with ≥1 CPT code for bariatric surgery and continuous enrollment in a health plan during the study period. Surgical procedures included gastric restriction without bypass performed with vertical banding (5.8%), gastric restriction without bypass performed with no vertical banding (4.1%), gastric restriction with bypass also known as short Roux-en-Y procedure (58.3%), gastric restriction with bypass, and small-bowel reconstruction (14.6%), laparoscopic restriction with bypass, also known as laparoscopic short Roux-en-Y (10.3%), laparoscopic with adjustable band (1.9%), via injection (2.0%), and laparoscopic gastric bypass (2.5%). Patient demographics and dropout rates are presented in Table 1. All the analyses presented below were also conducted on the sample of 1,372 patients with 3 years or more of follow-up with similar results.
Table 1

Patient baseline demographics


No. (%)

Study population (N)


Age (years, 95% CI)

44.3 (44.0, 44.5)


4,553 (82.8%)

Health plan type

Lock-in plans (∼health maintenance organization)

282 (5.1%)

Point of service

2,424 (44.1%)

Preferred provider organization

2,265 (41.2%)

Indemnity plans

282 (5.1%)



2,520 (45.8%)


859 (15.6%)


1,509 (27.4%)


614 (11.2%)

Sample retention

Censured because date of surgery was less than 3 years before Dec 2006

3,371 (61.3%)

Censured because individual dropped out of the sample within 3 years and before Dec 2006

759 (13.8%)

Individuals with more than 3 years of follow-up

1,372 (24.9%)

Changes in the prevalence of comorbidities between the presurgery period (days −90 to 0) and both the short-term (days 30 to 120) and medium-term (days 1,020 to 1,110) postsurgery periods are presented in Table 2 and over each 90-day interval in Fig. 1a–d.
Table 2

Comorbidity prevalence pre- and post-bariatric-surgery





Days −90 to 0

Days 30 to 120

Days 1,020 to 1,110

Number of patients with at least one claim (N = 5,502)

Number of patients with at least one claim (N = 5,502)

Number of patients with at least one claim (N = 1,372)

Obesity and other hyperalimentation


81.1% (83.1–82.2)

23.0% (21.9–24.2)*

4.4% (3.4–5.5)*

Cardiovascular disorders


43.6% (42.3–44.9)

23.4% (22.2–24.5)*

14.2% (12.4–16.1)*

 Hypertensive disease


38.3% (37.1–39.6)

21.1% (20.1–22.2)*

12.3% (19.6–14.1)*

 Cardiac dysrhythmias

426, 427

3.4% (3.5–3.9)

1.6% (1.2–1.9)*

1.5% (3.9–2.2)*

 Ischemic heart disease and cardiomyopathy


6.4% (5.8–7.1)

1.8% (1.5–2.2)*

1.9% (1.2–2.6)*

COPD and other respiratory conditions


57.7% (56.4–59.1)

26.1% (25.5–27.3)*

16.2% (14.2–18.1)*



7.6% (6.9–8.3)

3.0% (2.5–3.4)*

1.2% (4.7–1.8)*

 Sleeping disorders


30.5% (29.3–31.7)

19.8% (18.8–28.9)*

10.6% (9.4–12.3)*

Mental disorders


30.7% (29.5–31.9)

13.2% (12.3–14.1)*

14.8% (13.7–16.8)*

 Episodic mood disorders


7.7% (6.9–8.4)

5.9% (5.3–6.5)*

7.1% (5.8–8.5)

 Depressive disorder, not elsewhere classified


5.0% (4.4–5.6)

2.5% (2.1–3.6)*

2.8% (2.4–3.7)*

Diseases of the digestive system


39.4% (38.2–49.8)

26.7% (25.5–27.9)*

13.5% (11.7–15.3)*

Diseases of the musculoskeletal system and connective tissue


32.6% (31.4–33.9)

24.9% (23.8–26.1)*

27.7% (25.4–37.1)*

 Arthropathies and related disorders


16.6% (15.6–17.6)

9.8% (9.1–19.7)*

10.6% (8.9–12.2)*



12.8% (11.9–13.6)

10.2% (9.4–11.0)*

13.0% (11.2–14.8)

 Rheumatism, excluding the back


11.4% (13.6–12.3)

9.4% (8.7–19.3)*

10.6% (8.9–12.2)

 Osteopathies, chondropathies, and acquired musculoskeletal deformities


3.7% (3.2–4.2)

4.4% (3.9–4.9)

7.4% (6.3–8.8)*

Acute and chronic sinusitis, allergic rhinitis


9.7% (9.7–18.6)

6.3% (5.7–7.2)*

7.1% (5.7–8.4)*

Cerebrovascular disease


0.7% (2.5–1.7)

0.4% (0.2–0.6)*

0.5% (8.2–1.3)

Metabolic disorders

 Diabetes mellitus


19.9% (18.8–21.0)

15.0% (14.1–16.5)*

7.7% (6.3–9.1)*

 Disorders of lipoid metabolism


17.1% (16.1–18.1)

11.9% (11.1–12.8)*

5.7% (4.5–7.6)*



3.8% (3.3–4.3)

7.9% (7.2–8.7)*

9.9% (8.3–11.5)*

 Nutritional and mineral metabolism disorders

260–269, 275

1.0% (5.8–1.3)

7.1% (6.4–7.7)*

3.6% (2.6–4.6)*

Diseases of veins and lymphatics, and other diseases of circulatory system


3.6% (3.1–4.1)

3.4% (2.9–3.9)

2.6% (1.7–3.4)

*p < 0.05 (the difference between the presurgery and the postsurgery was statistically significant)

Fig. 1

a Short- and long-term effects of bariatric surgery on comorbidity claims—obesity and other hyperalimentation, hypertensive disease, ischemic heart disease, and cardiomyopathy, cardiovascular disorders. b Short- and long-term effects of bariatric surgery on comorbidity claims—asthma, sleeping disorders, COPD and other respiratory conditions, mental disorders. c Short- and long-term effects of bariatric surgery on comorbidity claims—diseases of the digestive system, diseases of the musculoskeletal system and connective tissue, diabetes mellitus, disorders of lipoid metabolism, acute and chronic sinusitis, allergic rhinitis. d Short- and long-term effects of bariatric surgery on comorbidity claims—anemia, nutritional and mineral metabolism disorders

Table 2 and Fig. 1a–d show a statistically and clinically significant decrease in claims associated with obesity and other hyperalimentation (71.6%; 95% confidence interval [CI] 70.2–73.0%) as early as 4 months following bariatric surgery. Similar statistically significant short-term decreases in claims associated with various comorbidities are observed, including claims for cardiovascular disorders (46.5%; CI 43.9–49.0%) as well as its components including hypertensive disease (45.0%; CI 42.2–47.8%), cardiac dysrhythmia (54.7%; 95%CI 45.2–64.2%), and ischemic heart disease (71.5%; CI 66.0–77.0%). Decreases in claims were also observed for chronic obstructive pulmonary disease (COPD) and other respiratory conditions (54.7%; CI 52.7–56.7%), including asthma (60.9%; CI 55.0–66.8%). Similar decreases in claims were found for sleep disorders (34.8%; CI 31.4–38.3%), mental disorders (56.9%; CI 54.0–59.8%), including episodic mood disorders (22.8%; CI 14.7–31.0%) and depressive disorders (48.9%; CI 40.6–57.2%), diseases of the digestive system (32.4%; CI 29.4–35.4%), diseases of the musculoskeletal system and connective tissues (23.7%; CI 20.2–27.2%), acute and chronic sinusitis, allergic rhinitis (35.3%; CI 28.7–41.9%), and cerebrovascular disease (45.0%; CI 22.1–67.9%). Among metabolic disorders, decreases in claims for diabetes mellitus (24.4%: CI 19.6–29.1%) and disorders of lipid metabolism (30.2%; CI 25.2–35.2%) were clinically and statistically significant. In contrast, increases in anemia claims (208.1%; CI 189.4–226.8%) and nutritional disorders (670.7%; CI 606.4–735.0%) were observed within 4 months of surgery. There was no statistically significant change in the prevalence of claims associated with diseases of veins and other diseases of the circulatory system, as these conditions were only sporadically present presurgery.

Table 2 and Fig. 1a–d demonstrate that the observed short-term decrease in comorbidity-associated claims was persistent over the 3-year period following surgery among the major disease categories except for claims associated with cerebrovascular disease, which first fell and then rose to a level statistically no different from the presurgery level but remained at a low absolute level (<1%). Similarly, some claims for specific disease categories, including claims for episodic mood disorders (7.1% 3 years following surgery versus 7.7% prior to surgery), dorsopathies (13.0% 3 years following surgery versus 12.8% prior to surgery), and rheumatism excluding the back (10.6% 3 years following surgery versus 11.4% prior to surgery), were not significantly different before and 3 years after surgery. Finally, there was no statistical change in claims associated with the prevalence of diseases of the veins and circulatory system over the 3-year period. The short-term increase in anemia claims (from 3.8% to 7.9%; p < 0.05) remained at 9.9% (p < 0.05) after 3 years. Claims associated with nutritional disorders which were rare prior to surgery (1.0% prevalence) increased to 7.1% within 4 months of surgery and then decreased to 3.6% after 3 years.

Comparisons in presurgery and postsurgery medication claims are reported in Table 3 and Fig. 2a, b by AHFS therapeutic class. Statistically significant decreases in medication claims (p < 0.05 for all comparisons) were observed within 4 months of surgery for several classes including anti-infectives (21.9%; CI 18.8–25.0%), psychotropics (13.0%; CI 9.7–16.3%), pain relievers (24.7%; CI 21.8–27.6%), respiratory drugs (37.3%; CI 32.6–41.9%), antihistamines (5.5%; CI −1.1% to 12.1%), hypolipidemics (55.3%; CI 50.8–59.7%), cardiovascular drugs (33.6%; CI 30.9–36.4%), and diabetic medications (including oral diabetic medication and insulin; 56.0%; CI 52.1–59.9%). There were no observed changes in claims for gastroenterology drugs (−0.6%; CI −4.6% to 3.4%) within 4 months of surgery. Claims for nutritional supplements increased during that same period (163.2%; CI 152.9–173.5%); however, because supplements are only sporadically reimbursed, claims are likely to significantly understate actual purchases and use.
Table 3

Pharmaceutical utilization pre- and post-bariatric-surgery





Days −90 to 0

Days 30 to 120

Days 1,020 to 1,110

Number of patients with at least one claim (N = 5,502)

Number of patients with at least one claim (N = 5,502)

Number of patients with at least one claim (N = 1,372)


08:08.00, 08:12.02, 08:12.04, 08:12.06,


08:12.12, 08:12.16, 08:12.18, 08:12.20,


08:12.24, 08:12.28, 08:14.04, 08:14.08,


08:14.28, 08:14.92, 08:18.32, 08:22.00,

39.4% (38.1–44.7)

30.8% (29.6–32.1)*

31.9% (29.5–34.4)*

08:24.00, 08:30.08, 08:30.92, 08:36.00,


08:40.00, 52:04.04, 52:04.08, 84:04.04,





28:16.08, 28:24.92, 28:24.08, 28:12.08,

37.4% (36.1–38.7)

32.6% (31.3–33.8)*

39.3% (36.7–41.9)

28:24.08, 28:28.00, 28:16.04


Pain reliever

28:08.04, 28:08.08, 28:08.12, 28:08.92,

42.3% (41.3–43.6)

31.8% (36.6–33.1)*

36.9% (34.3–39.4)

28:10.00, 52:08.00, 68:04.00, 84:06.00


Respiratory system

12:12.00, 12:16.00, 48:08.00, 48:16.00,

18.0% (17.5–19.1)

11.3% (12.5–12.2)*

10.9% (9.2–12.5)*

52:36.00, 68:04.00,



04:00.00, 04:04.00, 04:08.00, 52:02.00

13.2% (12.4–14.1)

12.5% (11.6–13.4)

10.2% (8.6–11.8)*


24:06.00, 24:06.04, 24:06.06, 24:06.08,

14.8% (13.9–15.7)

6.6% (6.2–7.3)*

5.3% (4.1–6.5)*




20:12.04, 24:04.04, 24:04.08, 24:08.00,


24:08.16, 24:12.08, 24:20.00, 24:24.00,

43.7% (42.4–45.3)

29.0% (27.8–33.2)*

25.5% (23.2–27.8)*

24:28.00, 24:28.08, 24:28.92, 24:32.04,


24:32.08, 24:32.20, 40:28.00, 40:28.10


Antidiabetic including

68:20.04, 68:20.08, 68:20.16, 68:20.20,

18.2% (17.2–19.2)

8.0% (7.3–8.7)*

5.1% (3.9–6.3)*

68:20.28, 68:20.92




4.9% (4.3–5.5)

2.3% (1.9–2.7)*

1.5% (3.9–2.2)*

Oral antidiabetics

68:20.04, 68:20.16, 68:20.20, 68:20.28,

16.3% (15.4–17.3)

6.6% (5.9–7.2)*

4.0% (3.8–5.1)*



Gastroenterologic including

12:08.08, 56:08.00, 56:12.00, 56:14.00,


56:22.00, 56:28.12, 56:28.32, 56:28.36,

30.6% (29.4–31.9)

30.8% (29.6–32.3)

16.1% (14.2–18.1)*

56:32.00, 56:40.00, 56:92.00



56:40.00, 56:28.36, 56:28.12, 56:40.00,

26.1% (25.2–27.3)

24.8% (23.7–26.8)

13.5% (11.7–15.3)*



Nutritional supplements

20:04.04, 40:12.00, 84:16.00, 88:08.00,

9.1% (8.3–9.8)

14.8% (13.9–15.8)*

12.6% (11.9–14.4)*



*p < 0.05 (the difference between the presurgery and the postsurgery was statistically significant)

Fig. 2

a Short- and long-term effects of bariatric surgery on medication utilization—anti-infective, psychotropic, pain reliever, cardiovascular, gastroenterologic, antacid. b Short- and long-term effects of bariatric surgery on medication utilization—respiratory system, hypolipidemic, antidiabetic, nutritional supplements, insulin, oral antidiabetics

For many of the AHFS therapeutic categories, gains achieved within 4 months were reinforced during the following 33 months (3 years postsurgery). Such persistence was observed for antihistaminics, cardiovascular treatment, diabetic medications, and gastroenterologic drugs. Decreases for anti-infectives, pain relievers, and respiratory system treatment were partially reversed over the following 33 months. Claims for psychotropic drugs were not significantly different after 3 years compared to presurgery levels despite the decrease observed at 4 months. Finally, with a note of caution because of sporadic reimbursement, claims for nutritional supplements seem to have decreased relative to their level at 4 months but remained higher than prior to surgery.


Observed reductions in claims for major comorbidities as well as reduced medication use suggest a rapid and systemic improvement in the health profiles of patients likely resulting from the rapid weight loss associated with bariatric surgery. The kinetics of the prevalence of comorbidities suggests an immediate decrease in comorbidities (within 4 months of surgery) that is maintained over the 3-year period of observation.

Our findings reinforce existing literature on the impact of bariatric surgery on major comorbidities. Three years after surgery, the decrease in prevalence for diabetes (61%), hypertension (68%), and disorders of lipid metabolism (66%) is close to those reported in recent meta-analyses [10, 11]. However, these results contrast with those of Zingmond et al. who found significant increases in hospitalizations following bariatric surgery (from 7.9% to 19.3%) [11]. The observed increases resulted mainly from hernia repair and revisional gastric bypass surgery and may not be inconsistent with reduced comorbidities for the general bariatric population.

This study focused on the 50 most prevalent conditions observed prior to surgery. This approach highlights the high prevalence of several diseases in morbidly obese patients undergoing bariatric surgery including psychiatric, respiratory, rheumatoid, cardiac, and gastrointestinal disorders, all of which decreased to an extent consistent with prior research [6]. The observed decrease in claims associated with these comorbidities suggests that the presurgery conditions resulted from patients’ excess weight and is alleviated by bariatric surgery for at least 3 years.

The level and change in prevalence of psychiatric disorders are difficult to interpret because the screening for psychiatric disorder often performed to determine eligibility for bariatric surgery is typically accompanied by mental health treatment and further screening initiatives that are likely to bias any presurgery and postsurgery comparisons. The same caveat applies to psychotropic agents. Notwithstanding the caveats outlined above, the decrease in the prevalence of claims for psychiatric disorders after surgery is consistent with prior results [6].

In contrast with prior research which suggested a significantly higher incidence of gastrointestinal disorders 5 years after bariatric surgery compared to a control group (36.4% vs. 24.7%; p < 0.05) [6], our research shows that diseases of the digestive system decreased significantly up to 3 years postsurgery. This discrepancy could result from the experimental design differences, different observation lengths (3 years versus 5 years), or on the control group selection in Christou et al., not at issue in our longitudinal design [6].

Chronic complications of bariatric surgery include nutritional deficiencies, including anemia-inducing reduced iron, vitamin B12, or folates, which are insufficiently evaluated [12, 13]. We find an increase in the prevalence of anemia from 3.8% to 9.9% within 3 years consistent with a study on a small sample of patients reporting lower hemoglobin levels after gastric bypass [12].

The dramatic decrease in the prevalence of major diseases is paralleled with reduced medication use related to major comorbidities such as cardiovascular disease, diabetes, and hyperlipidemia in both the short-term and medium-term postsurgical periods. The reduced use of diabetes and cardiovascular medications and increased use of nutritional supplements is consistent with prior findings [14], while the statistically significant decrease in the prevalence of pain relievers, anti-infectives, and respiratory medications after surgery has not been reported previously. Our finding of an approximately 50% decrease in the use of gastrointestinal medication 3 years after surgery contrasts with prior findings of higher use among bariatric patients and may be related to a decreased incidence of gastroesophageal reflux disease following weight loss. There is little literature on the effect of bariatric surgery on the use of or claims for psychotropic medication. However, our results depart from prior findings which suggested reduced antidepressant use at 12 months [14].

To our knowledge, the early kinetics of the prevalence of comorbidities and medication use following bariatric surgery has never been reported before. The parallel and persistent decreases in disease prevalence and related medication use as early as 4 months after surgery have important medical implications and demonstrate the immediate benefits of bariatric surgery on health status. They also suggest that if improvement in comorbidities is not observed by 4 months the likelihood of a later improvement is low.

Our study has several limitations. The analysis was based on a retrospective review of claims and is not a randomized, prospective design. It offers an exhaustive record of diagnoses and medications that reflect actual clinical use and have been reported to have high specificity but varying sensitivity across comorbidities but do not include clinical data [15]. The reduced prevalence of comorbidities is inferred from a reduced level of claims rather than any direct clinical observation. Similarly, the reduced level of medication claims reflects reduced purchases of medication from which reduced use can only be inferred. Finally, uninsured care or uninsured medication purchases (e.g., over the counter) are not recorded in the data used here.

As in other studies, the patient population who underwent bariatric surgery is over 80% women, limiting the applicability of our findings to men [9, 11, 16]. Furthermore, despite differences in obesity prevalence across racial groups [1], the lack of data on race made between-group analyses impossible. Finally, although close to 60% of the bariatric procedures were of one specific type (gastric restriction with bypass; short Roux-en-Y), the procedures were heterogeneous and were not analyzed separately because of coding limitations. European studies, such as the SOS intervention study, reported gastric restriction without bypass as the most common type of surgery; hence, our results should not be extrapolated to a European population particularly since prior analyses have demonstrated differential effectiveness across the two types of procedures [10, 11, 14].

The decrease in statistical power associated with the small percentage of patients (25%) for whom 3 years of postsurgery data were observed is of limited concern because of the large absolute sample size (n = 1,372). However, to the extent that more recent surgeries are better targeted, accompanied by better screening and patient support, or are more effective than earlier interventions, our analysis may understate the reduction in comorbidities and medication use.

Financial Disclosure

The authors of this paper did not receive any funding for the analysis. Pierre Cremieux and Marric Buessing received funding for an analysis of the return on investment of bariatric surgery funded by Ethicon Endo-Surgery, Inc., and published in the September 2008 issue of the American Journal of Managed Care under the title, “A Study on the Economic Impact of Bariatric Surgery.”

Copyright information

© Springer Science + Business Media, LLC 2010