Abstract
Aims
Obesity is a risk factor for multiple long‐term conditions (MLTCs/multimorbidity). However, the impact of weight loss in people with MLTCs is unclear. We investigated the association between body mass index (BMI) change and the development of obesity‐related complications (ORCs), as well as clinical and economic outcomes in individuals with obesity and MLTCs.
Materials and Methods
This cohort study included adults aged ≤70 years with BMI ≥30 kg/m2 and ≥2 ORCs. BMI was recorded during Years 1 and 4 of the baseline period. BMI change was categorised as increases or decreases of 3%–7%, 7%–15% and 15%–30%. Data were analysed using Cox regression (hazard ratios [HRs] for mental‐health conditions, hospitalisation and mortality) and Ghosh–Lin models (risk of ≥1 new ORC and clinical consultations).
Results
A total of 618 426 individuals were included. The mean number of ORCs at index (Year 4) was 3.8 (standard deviation 1.5). HRs (95% confidence intervals) for incident ORCs were 0.96 (0.94–0.98), 0.98 (0.96–0.99) and 0.98 (0.97–0.99) for the highest to lowest BMI reductions, respectively; BMI increases were associated with HRs >1.00. Risk ratios for consultations were 0.99 (0.98–1.00), 0.99 (0.98–1.00) and 1.00 (0.99–1.00). BMI reductions were linked to lower polypharmacy rates. Both BMI decrease and increase were associated with higher HRs for mental‐health conditions, hospitalisation and mortality versus stable BMI.
Conclusions
Weight loss in individuals with obesity and MLTCs was linked to both favourable and adverse outcomes, highlighting the importance of personalised treatment approaches that consider outcomes beyond weight loss.
Keywords: BMI, multiple long‐term conditions, obesity, obesity‐related complications, weight gain, weight loss
1. INTRODUCTION
Multiple long‐term conditions (MLTCs; also known as multimorbidity), most commonly defined as the occurrence of ≥2 chronic conditions in an individual, have adverse impacts on providers, patients and health care systems. 1 , 2 The presence of MLTCs is associated with an increased risk for suboptimal care, 3 , 4 poorer quality of life and impaired mental health, 2 , 5 more frequent and longer hospital stays, higher health care costs, and increased polypharmacy and mortality. 6 , 7 , 8 The true prevalence of MLTCs is difficult to determine, as the definition and measurement of MLTCs vary by study; however, a systematic review of 126 peer‐reviewed studies proposed a global prevalence estimate of 37.2% (95% confidence interval [CI] 34.9–39.4) in the general population. 9 Prevalence estimates of MLTCs rise with increasing age and increasing level of socioeconomic deprivation. 9 , 10
Alongside ethnicity, smoking, diabetes, sedentary behaviour and socioeconomic deprivation, obesity is a major risk factor in the development of MLTCs. 11 , 12 People with obesity have two‐fold‐higher odds of having MLTCs than people without obesity. 8 , 13 In a pooled analysis of 16 prospective cohort studies, the odds of developing cardiometabolic MLTCs were 4.5 times higher among those with class I obesity (body mass index [BMI] 30–35 kg/m2) and 14.5 times higher among those with class III obesity (BMI ≥35 kg/m2) versus those with BMI 20 to <25 kg/m2. 11 People with obesity (BMI ≥30 kg/m2) have been shown to have five times the risk of developing simple MLTCs (two coexisting diseases) and 12 times the risk of complex MLTCs (four coexisting diseases) versus those who have a BMI of 18.5 to <25 kg/m2. 14
Moderate weight loss and maintenance through dietary and lifestyle changes, pharmacological intervention or metabolic surgery have been shown to reduce the disease burden, healthcare resource utilisation (HCRU), and costs by reducing the risk of obesity‐related conditions (ORCs). 15 , 16 , 17 , 18 Although some studies have addressed the impact of weight reduction on ORCs, there is little information on the impact of obesity management and weight change in individuals with MLTCs.
The aim of this study was to assess the association between BMI change and (1) the risk of developing ORCs and mental‐health conditions, (2) mortality, (3) HCRU and (4) polypharmacy in individuals with obesity and MLTCs. Considering the relationship between obesity and MLTCs, we hypothesised that weight loss would have favourable outcomes in those with MLTCs.
2. MATERIALS AND METHODS
2.1. Data source
This was a cohort study on data from the period 1 January 2001 to 31 January 2022 from the Clinical Practice Research Datalink (CPRD) Aurum database. CPRD Aurum contains electronic medical records of >60 million patients seen by general practitioners (GPs) in the UK, including >18 million current patients. 19 Data from CPRD Aurum were linked by NHS England to (1) Office for National Statistics mortality data; (2) Hospital Episode Statistics (HES) data on admitted patient, outpatient, and accident and emergency care; and (3) the Index of Multiple Deprivation (IMD). Patient consent was not required, as data sent to CPRD Aurum by GPs did not contain identifiable, private health information. Approval from an Independent Scientific Advisory Committee was obtained (reference #23_002635).
2.2. Study design and patient population
The study included a 5‐year screening period, 4‐year baseline period (during which BMI change was assessed) and 4‐year follow‐up period (Figure 1). Eligibility was assessed during the 5‐year screening period, and individuals were required to have ≥2 ORCs, be aged ≥18 to ≤70 years, and have received their first obesity diagnosis (defined as BMI ≥30 or ≥25 kg/m2 for Asian individuals) between 1 January 2001 and 31 December 2017. The date of obesity diagnosis signified the start of the baseline period. Individuals were also required to have a BMI measurement in Year 4 of the baseline period to determine BMI change. The date of the last BMI measurement in Year 4 of the baseline period was the index date and signified the start of the follow‐up period. Change in BMI was recorded as the difference between the average BMI in Year 1 of the baseline and the average BMI in Year 4 (index), expressed as a percentage. BMI change was categorised as stable (within ≥−3% to <3%), increased (increases of ≥3% to <7%, ≥7% to <15% or ≥15% to <30%) or decreased (decreases of <−3% to ≥−7%, <−7% to ≥−15% or <−15% to >−30%). If multiple measurements were recorded during Year 4 of the baseline period, the average BMI for that year was calculated. Individuals were followed from the index date, defined as the date of the last BMI measurement at any time in Year 4, for up to 4 years until either the end of the study period, the end of the individual's record or the last date of information available from CPRD Aurum.
FIGURE 1.
Study design. BMI, body mass index; IMD, Index of Multiple Deprivation; ORC, obesity‐related complication.
Individuals were excluded if they developed malignant cancer, thyroid disorder, cachexia, heart failure with reduced ejection fraction or dementia during the screening period, or if they had a record for pregnancy or limb amputation during the baseline period.
2.3. Study variables
This study assessed the association between BMI change and (1) the hazard of developing a new ORC, (2) the risk of developing ≥1 recurrent ORC, (3) the hazard of developing a mental‐health condition, (4) mortality, (5) the occurrence of clinical consultations and hospitalisation, and (6) the rate of polypharmacy.
Demographic data were collected at the time of obesity diagnosis and included sex, age, ethnicity, BMI, obesity class, smoking status, alcoholism (diagnosis of alcohol misuse or any alcohol‐related disorder), prevalent ORCs and IMD to define socioeconomic status (based on the individual's residence, with IMD quintile 1 representing the least deprived group; individuals with missing IMD data were categorised as 0). BMI was also collected during Year 4 of the baseline period.
Smoking status was recorded by the physician (GP) as part of routine consultations with patients, with the majority held face to face as part of a routine review. Patients would normally be asked about smoking history and would be coded in electronic health records as a previous or current or non‐smoker; if the patient was a current smoker, they would be asked how many cigarettes they smoked per day.
2.3.1. Obesity‐related conditions
Two analyses were conducted on ORCs by BMI change, including a time‐to‐first‐event analysis investigating the hazard of developing one additional new ORC and a recurrent event analysis investigating the risk of developing ≥1 ORC (i.e., the risk of having a greater disease count). The list of ORCs was chosen with guidance from Kivimäki et al. 14 and Ho et al. 20 ORCs were identified during the baseline and follow‐up periods from diagnosis codes in CPRD Aurum or International Classification of Diseases, Tenth Revision (ICD‐10) codes in the HES data (Appendix 2, Supporting Information S1), and included asthma, back pain, chronic obstructive pulmonary disease, chronic kidney disease, deep vein thrombosis, dementia, eczema, dyslipidaemia, gout, gastroesophageal reflux disease, heart failure, hypertension, ischaemic heart disease, musculoskeletal pain (e.g., fibromyalgia), osteoarthritis (e.g., hip and knee), obesity‐related cancers, polycystic ovarian syndrome, peripheral vascular disease, pre‐diabetes, metabolic‐associated fatty liver disease/metabolic‐associated steatohepatitis, psoriasis, pulmonary embolism, renal failure, rheumatoid arthritis and related disorders, sleep disorders (e.g., insomnia and obstructive sleep apnoea), stroke, transient ischaemic attack, type 2 diabetes and urinary incontinence in females. Individuals who had a previous record for a particular ORC during the baseline or follow‐up periods were not considered at risk for developing that condition again. Both the incidence of a new ORC and count of prevalent ORCs were analysed.
2.3.2. Mental‐health conditions
The incidence of mental‐health conditions was identified during the baseline and follow‐up periods via diagnosis codes in CPRD Aurum or ICD‐10 codes in the HES data, and included depression, anxiety and self‐harm. Individuals who had a previous record for these conditions were included to determine whether the findings also applied to a population at increased risk of mental‐health disorders. A time‐to‐first‐event analysis was used to determine the risk of developing a mental‐health condition, analysed by BMI change. Individuals who had a previous record for a particular mental‐health condition during the baseline or follow‐up periods were not considered at risk for developing that condition again and were considered for other outcomes.
2.3.3. Mortality
The number of all‐cause deaths was measured during follow‐up. The hazard of mortality, analysed by BMI change, was investigated using a time‐to‐first‐event analysis.
2.3.4. Healthcare resource utilisation
Occurrence and risk of a recurrent clinical consultation and time to first hospitalisation were measured during the baseline and follow‐up periods and analysed by BMI change. Clinical consultations were identified via diagnosis codes in CPRD Aurum (Appendix 1, Supporting Information: Table 1) and included face‐to‐face, remote and telephone consultations with a nurse or GP.
2.3.5. Polypharmacy
Medication records quantities were compared at the time of diagnosis and during follow‐up. Change in the number of records over time was analysed by BMI change. All medication records, regardless of indication or administration type, were included in the analysis.
2.4. Statistical analysis
Statistical analyses were conducted using RStudio (Survival Biostats group) and analysis of multivariate events times and survival modelling packages. Hazard ratios (HRs) were computed using Cox proportional hazard models implemented with the ‘coxph’ function from the R‐package ‘survival’ package version 3.8‐3. 21 Risk (count) ratios were computed using Ghosh–Lin models implemented with the ‘recreg’ function from the R‐package ‘mets’ version 1.3.4. 22
Cox proportional hazards models were used for time‐to‐first‐event analyses, including the hazard of developing an additional ORC, developing a mental‐health condition, hospitalisation and mortality by percent change in BMI. All models were adjusted for index year, calendar year, age, sex, ethnicity, alcoholism, smoking status and number of ORCs at baseline. Models analysing the association between BMI change and the hazard of developing an additional ORC, developing a mental‐health condition and mortality were also adjusted for IMD.
Recurrent event analyses, including the association between BMI change and the risk of developing ≥1 ORC and the occurrence of clinical consultations were assessed via Ghosh–Lin inverse probability of censoring weighted models. The inverse probability of censoring weights were incorporated, providing unbiased estimates of the covariate effects in the presence of informative censoring. The models were adjusted for index year, age, sex, ethnicity, alcoholism, smoking status, number of ORCs at baseline and IMD.
The relationship between the percent change in BMI and change in polypharmacy was measured during the BMI change period and modelled using a repeated measures model.
Change in BMI (expressed as a percentage) reflected an average proportional difference (i.e., capturing the overall linear change in BMI).
Additional sensitivity analyses were conducted for all endpoints (Appendix 1, Supporting Information: Methods).
3. RESULTS
3.1. Demographic characteristics
A total of 618 426 individuals met the inclusion criteria and were included in the analyses. At diagnosis, 50.4% were male, and most individuals (63.7%) had class I obesity (Table 1). Mean (standard deviation [SD]) age was 52.0 (12.0) years at index (Year 4). Mean (SD) number of ORCs was 2.0 (1.0) at diagnosis and 3.8 (1.5) at index (Year 4). Mean (SD) BMI was 34.2 (4.7) kg/m2 at diagnosis and 34.3 (5.2) kg/m2 at index. During the 4‐year baseline period, 29.3% of individuals had a BMI decrease, 37.6% had stable BMI and 33.1% had a BMI increase (Table 1).
TABLE 1.
Patient characteristics during the baseline period.
| Percent change in BMI from Year 1 to Year 4 of the baseline period | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Decrease | Stable | Increase | |||||||
| Overall at diagnosis (n = 618 426) | >−30% to <−15% (n = 18 970) | ≥−15% to <−7% (n = 65 963) | ≥−7% to <−3% (n = 96 196) | ≥−3% to <3% (n = 232 528) | ≥3% to <7% (n = 110 233) | ≥7% to <15% (n = 77 468) | ≥15% to <30% (n = 17 068) | p value | |
| Index year a | <0.001 | ||||||||
| 2001–2005 | 202 455 (32.7) | 5198 (27.4) | 19 347 (29.3) | 29 837 (31.0) | 76 371 (32.8) | 38 606 (35.0) | 27 492 (35.5) | 5604 (32.8) | |
| 2006–2010 | 251 587 (40.7) | 7975 (42.0) | 27 316 (41.4) | 39 301 (40.9) | 95 492 (41.1) | 44 204 (40.1) | 30 468 (39.3) | 6831 (40.0) | |
| 2011–2015 | 164 384 (26.6) | 5797 (30.6) | 19 300 (29.3) | 27 058 (28.1) | 60 665 (26.1) | 27 423 (24.9) | 19 508 (25.2) | 4633 (27.1) | |
| Sex b | <0.001 | ||||||||
| Female | 306 892 (49.6) | 12 064 (63.6) | 35 909 (54.4) | 46 341 (48.2) | 106 054 (45.6) | 53 411 (48.5) | 42 443 (54.8) | 10 670 (62.5) | |
| Male | 311 534 (50.4) | 6906 (36.4) | 30 054 (45.6) | 49 855 (51.8) | 126 474 (54.4) | 56 822 (51.5) | 35 025 (45.2) | 6398 (37.5) | |
| Age at index a | 52.0 (12.0) | 49.1 (13.8) | 51.7 (12.5) | 53.2 (11.6) | 53.3 (11.3) | 51.9 (11.7) | 49.3 (12.7) | 44.9 (13.9) | <0.001 |
| Ethnicity b | <0.001 | ||||||||
| Asian | 50 190 (8.1) | 823 (4.3) | 4134 (6.3) | 7206 (7.5) | 19 151 (8.2) | 9902 (9.0) | 7357 (9.5) | 1617 (9.5) | |
| Non‐Asian | 568 236 (91.9) | 18 147 (95.7) | 61 829 (93.7) | 88 990 (92.5) | 213 377 (91.8) | 100 331 (91.0) | 70 111 (90.5) | 15 451 (90.5) | |
| Substance abuse b | <0.001 | ||||||||
| Yes | 18 083 (2.9) | 765 (4.0) | 2237 (3.4) | 2703 (2.8) | 6452 (2.8) | 3106 (2.8) | 2244 (2.9) | 576 (3.4) | |
| No | 600 343 (97.1) | 18 205 (96.0) | 63 726 (96.6) | 93 493 (97.2) | 226 076 (97.2) | 107 127 (97.2) | 75 224 (97.1) | 16 492 (96.6) | |
| Smoking b | <0.001 | ||||||||
| Yes | 118 639 (19.2) | 5051 (26.6) | 15 517 (23.5) | 18 549 (19.3) | 39 503 (17.0) | 19 432 (17.6) | 16 107 (20.8) | 4480 (26.2) | |
| No | 499 787 (80.8) | 13 919 (73.4) | 50 446 (76.5) | 77 647 (80.7) | 193 025 (83.0) | 90 801 (82.4) | 61 361 (79.2) | 12 588 (73.8) | |
| Number of ORCs at index a | 3.8 (1.5) | 3.7 (1.6) | 3.9 (1.6) | 3.9 (1.5) | 3.8 (1.5) | 3.8 (1.5) | 3.8 (1.6) | 3.8 (1.6) | <0.001 |
| Number of ORCs a | n = 18 996 | n = 65 945 | n = 96 205 | n = 232 532 | n = 110 222 | n = 77 468 | n = 17 058 | ||
| 2 | 124 310 (20.1) | 4603 (24.2) | 13 196 (20.0) | 17 732 (18.4) | 45 214 (19.4) | 22 260 (20.5) | 16 842 (21.7) | 4103 (24.0) | |
| 3 | 174 376 (28.2) | 5296 (27.9) | 18 046 (27.4) | 26 839 (27.9) | 66 311 (28.5) | 31 454 (28.5) | 21 794 (28.1) | 4636 (27.2) | |
| 4 | 145 944 (23.6) | 4047 (21.3) | 15 312 (23.2) | 23 556 (24.5) | 55 935 (24.1) | 25 990 (23.6) | 17 530 (22.6) | 3574 (21.0) | |
| 5+ | 173 796 (28.1) | 5050 (26.6) | 19 391 (29.4) | 28 078 (29.1) | 65 072 (28.0) | 30 158 (27.4) | 21 302 (27.5) | 4745 (27.8) | |
| Number of mental‐health conditions a | 0.3 (0.6) | 0.4 (0.7) | 0.3 (0.6) | 0.3 (0.6) | 0.2 (0.6) | 0.3 (0.6) | 0.3 (0.7) | 0.5 (0.8) | <0.001 |
| IMD b | <0.001 | ||||||||
| 0 | 91 621 (14.8) | 2859 (15.1) | 9882 (15.0) | 14 268 (14.8) | 34 613 (14.9) | 16 204 (14.7) | 11 303 (14.6) | 2492 (14.6) | |
| 1 | 80 211 (13.0) | 2294 (12.1) | 8425 (12.8) | 12 844 (13.4) | 31 720 (13.6) | 14 153 (12.8) | 9091 (11.7) | 1684 (9.9) | |
| 2 | 94 447 (15.3) | 2791 (14.7) | 9666 (14.7) | 15 043 (15.6) | 36 696 (15.8) | 16 885 (15.3) | 11 088 (14.3) | 2278 (13.3) | |
| 3 | 98 586 (15.9) | 2992 (15.8) | 10 560 (16.0) | 15 223 (15.8) | 37 431 (16.1) | 17 739 (16.1) | 12 087 (15.6) | 2554 (15.0) | |
| 4 | 117 851 (19.1) | 3530 (18.6) | 12 531 (19.0) | 18 360 (19.1) | 43 667 (18.8) | 21 093 (19.1) | 15 214 (19.6) | 3456 (20.2) | |
| 5 | 135 710 (21.9) | 4504 (23.7) | 14 899 (22.6) | 20 458 (21.3) | 48 401 (20.8) | 24 159 (21.9) | 18 685 (24.1) | 4604 (27.0) | |
| BMI during Year 1 of the baseline period | 34.2 (4.7) | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | |
| BMI during Year 4 of the baseline period | 34.3 (5.2) | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | ‐ | |
| Percent change in BMI c | 0.23 (7.5) | −19.8 (3.9) | −10.0 (2.2) | −4.8 (1.1) | 0.1 (1.7) | 4.8 (1.1) | 9.9 (2.2) | 19.2 (3.6) | <0.001 |
| Obesity class d | <0.001 | ||||||||
| I | 394 091 (63.7) | 9161 (48.3) | 38 244 (58.0) | 60 601 (63.0) | 153 456 (66.0) | 72 266 (65.6) | 49 460 (63.8) | 10 903 (63.9) | |
| II | 143 184 (23.2) | 4835 (25.5) | 16 096 (24.4) | 22 473 (23.4) | 52 508 (22.6) | 25 027 (22.7) | 18 180 (23.5) | 4065 (23.8) | |
| III | 81 151 (13.1) | 4974 (26.2) | 11 623 (17.6) | 13 122 (13.6) | 26 564 (11.4) | 12 940 (11.7) | 9828 (12.7) | 2100 (12.3) | |
| Medication use b | <0.001 | ||||||||
| Antihypertensive | 381 126 (61.6) | 10 337 (54.5) | 39 200 (59.4) | 59 926 (62.3) | 147 545 (63.5) | 68 890 (62.5) | 46 078 (59.5) | 9150 (53.6) | |
| Weight loss drug | 8134 (1.3) | 314 (1.7) | 1309 (2.0) | 1627 (1.7) | 2771 (1.2) | 1092 (1.0) | 808 (1.0) | 213 (1.2) | |
| Anticoagulant | 32 532 (5.3) | 1359 (7.2) | 3946 (6.0) | 4912 (5.1) | 11 474 (4.9) | 5532 (5.0) | 4248 (5.5) | 1061 (6.2) | |
Note: Data are n (%) or mean (SD).
Abbreviations: BMI, body mass index; IMD, Index of Multiple Deprivation; ORC, obesity‐related complication; SD, standard deviation.
Assessments were at the index date (the date of the BMI measurement in Year 4, and start of the follow‐up period).
Assessments were at first obesity diagnosis (the start of the baseline period).
BMI change was between the date of the first eligible BMI record (Year 1 of the 4‐year baseline period) and the date of the eligible BMI measurement in Year 4 of the baseline period (the date of the last BMI measurement in Year 4 was the index date).
Obesity based on first BMI diagnosis and measurement at the start of the baseline period: class I BMI = 30 to <35 kg/m2; class II BMI = 35 to <40 kg/m2; class III BMI = ≥40 kg/m2.
3.2. Hazard of developing an additional obesity‐related complication
During follow‐up, 340 775 new ORCs were observed among 1.4 million person‐years. Compared with individuals with stable BMI, those whose BMI increased over the baseline period had a significantly higher adjusted HR of developing one additional ORC, whereas individuals whose BMI decreased had a significantly lower adjusted HR (Figure 2A).
FIGURE 2.
Association between percent change in body mass index (BMI) during the baseline period and (A) hazard of developing an additional obesity‐related complication (ORC) and (B) risk of developing ≥1 ORC, both during follow‐up. The hazard ratio (HR) of an additional ORC was analysed with a Cox proportional hazards model using stable weight (BMI change −3% to 3%) as the reference. The risk ratio of developing ≥1 ORC was analysed with a Ghosh–Lin probability of censoring weighted model. Both models were adjusted for index year, age, sex, ethnicity, alcoholism, smoking status, Index of Multiple Deprivation and the number of ORCs at baseline. Numbers are presented as HR or risk ratio (95% confidence interval [CI]). Data are plotted to each exact value, and data labels have been rounded to two decimal points.
Female sex, Asian ethnicity, alcoholism, smoking, a greater level of socioeconomic deprivation, age and number of ORCs at baseline were also associated with a higher hazard of developing one additional ORC (Appendix 1, Supporting Information: Figure 1).
3.3. Risk of developing ≥1 obesity‐related complication (disease count)
During follow‐up, 55.0%, 26.7% and 11.0% of individuals developed an additional one, two and three ORCs, respectively, compared with those reported at index. Association between BMI change and the risk of developing ≥1 ORC followed a similar pattern as noted above; compared with individuals with stable BMI, the relative risk of developing ≥1 ORC during follow‐up was higher among individuals whose BMI increased during the baseline period and lower among those with decreased BMI (Figure 2B).
3.4. Risk of mental‐health conditions
A total of 8.9% of the study population (n = 54 914) had a record for a mental‐health condition (Appendix 1, Supporting Information: Figure 2) during follow‐up. Relative risk of being diagnosed with a mental‐health condition increased significantly among individuals who had a BMI increase of ≥7% or a BMI decrease compared with those who had stable BMI during the baseline period (Figure 3A). Female sex, not being of Asian ethnicity, alcoholism, smoking, age, number of ORCs in the baseline period, and being in the highest IMD quintile were associated with a greater risk of mental‐health diagnoses during follow‐up (Appendix 1, Supporting Information: Figure 3A).
FIGURE 3.
Association between percent change in body mass index (BMI) during the baseline period and (A) hazard of mental‐health conditions and (B) hazard of mortality, both during follow‐up. The relationships between percent change in BMI and the hazards of mental‐health conditions (i.e., anxiety, depression or self‐harm) and mortality were analysed with a Cox proportional hazards model using stable weight (BMI change −3% to 3%) as the reference. Both models were adjusted for index year, age, sex, ethnicity, alcoholism, smoking status, Index of Multiple Deprivation and number of obesity‐related complications at baseline. Numbers are presented as hazard ratio (HR) (95% confidence interval [CI]). Data are plotted to each exact value, and data labels have been rounded to two decimal points.
3.5. Hazard of mortality
A total of 15 564 deaths were observed among 2.2 million person‐years during follow‐up. The HR for mortality was significantly higher among those with decreased and increased BMI compared with those who had stable BMI during the baseline period (Figure 3B). The hazard of mortality was 3.3 times higher among those with the greatest BMI decrease and 2.3 times higher among those with the greatest BMI increase versus those with stable BMI (p < 0.001 for both comparisons). A higher mortality HR was associated with age, male sex, not being of Asian ethnicity, alcoholism, smoking, greater socioeconomic deprivation and number of ORCs during the baseline period (Appendix 1, Supporting Information: Figure 3B). When stratifying by obesity class, the U‐shaped curve for the hazard of mortality flattened with increasing obesity class (Appendix 1, Supporting Information: Figure 4).
3.6. Healthcare resource utilisation
A total of 87.1% of individuals (n = 538 595) had a record for a clinical consultation (including emergency appointments/urgent visits) during follow‐up. The occurrence of clinical consultations among those with ≥1 record is shown in Appendix 1, Supporting Information: Figure 5A. Individuals with BMI reductions ≥7% had a significantly lower occurrence compared with those with stable BMI (Figure 4A). The risk ratio (95% CI) for a clinical consultation was 1.01 (1.00–1.03) among those with the greatest BMI increase and 0.99 (0.98–1.00) among those with the greatest BMI decrease.
FIGURE 4.
Association between percent change in body mass index (BMI) during the baseline period and (A) occurrence of clinical consultations, (B) hazard of hospitalisation and (C) polypharmacy change, all during follow‐up. The relationships between percent change in BMI and (A) occurrence of clinical consultations (face‐to‐face and remote consultations with a nurse or general practitioner) and (B) hazard of hospitalisation were analysed with a Ghosh‐Lin probability of censoring weighted model and a Cox proportional hazards model, respectively, using stable weight (BMI change −3% to 3%) as the reference. The models were adjusted for index year, age, sex, ethnicity, alcoholism, smoking status and number of obesity‐related complications at baseline. The Cox proportional hazards model was also adjusted for Index of Multiple Deprivation. Numbers are presented as risk ratio (95% confidence interval [CI]) and hazard ratio (HR) (95% CI). (C) The relationship between percent change in BMI and mean change in polypharmacy (change in the number of medication records) was analysed via a repeated measures model. Numbers are estimated mean changes (95% CI). Data are plotted to each exact value, and data labels have been rounded to two decimal points.
During follow‐up, 12.8% of individuals (n = 79 061) had a record for hospitalisation. The distribution of the number of hospitalisations among those with ≥1 is shown in Appendix 1, Supporting Information: Figure 5B. The HR for hospitalisation was significantly higher among those with decreased and increased BMI compared with those with stable BMI (Figure 4B). Age, female sex, Asian ethnicity, alcoholism, smoking and number of ORCs in the baseline period were associated with a greater hazard of hospitalisation during follow‐up (Appendix 1, Supporting Information: Figure 6).
3.7. Polypharmacy
Individuals with stable BMI during the baseline period had a mean polypharmacy increase of 21.2% during follow‐up (Figure 4C). Individuals with a BMI decrease of ≥7% during follow‐up had a significantly lower increase in polypharmacy from index compared with those with stable BMI, with an increase of 18.8% in those with the greatest BMI decrease. In contrast, individuals with a BMI increase of ≥7% had a significantly higher increase in polypharmacy from index compared with those with stable BMI, with a 22.8% increase in those with the greatest BMI increase.
3.8. Sensitivity analyses
Results from the sensitivity analyses (shown in Appendix 1, Supporting Information: Methods) were consistent with the main results.
4. DISCUSSION
In a population with high disease burden due to obesity and MLTCs, weight loss was associated with fewer clinical consultations (face‐to‐face and remote) and the development of fewer ORCs compared with individuals with stable weight over the 4‐year follow‐up period. Weight gain was associated with higher rates of clinical consultations and the development of a higher number of ORCs. However, both weight loss and weight gain were associated with an increased hazard of mental‐health conditions and hospitalisation, and a greater mortality risk, with all showing a U‐shaped curve from weight loss to weight gain. Although all individuals had an increase in polypharmacy from index (which is expected with increasing age during the long follow‐up period), those with the greatest weight loss experienced the lowest increase, and those with the greatest weight gain experienced the largest increase. Although BMI change captured average proportional differences rather than a detailed longitudinal trajectory, the target population in this study included both non‐linear and linear BMI fluctuations. Thus, our results are relevant for both subgroups, regardless of how BMI changed dynamically.
Weight loss had a favourable impact on development of ORCs in this study. A dose–response relationship was observed between percent change in BMI and both the HR of developing one additional ORC and the risk ratio of developing multiple ORCs. The greater the BMI reduction, the lower the hazard of developing an additional ORC, and the lower the risk of developing ≥1 ORC. The reverse was also true: the greater the BMI increase, the higher the hazard of developing an additional ORC, and the higher the risk of developing ≥1 ORC. This is consistent with previous observational and interventional studies that evaluated lifestyle interventions, pharmacotherapy or metabolic surgery. 23 , 24 , 25 , 26 , 27 This study extended the findings of these studies to a population with high disease burden at baseline, showing that weight change affects risk of ORCs, even in a population with MLTCs and high ORC burden at baseline. Therefore, treating obesity should be considered as an option for alleviating risk of ORCs in people with MLTCs and obesity.
Weight loss in this study was also associated with lower polypharmacy rates compared with stable weight. Polypharmacy is associated with various adverse health outcomes and can lead to reduced quality of life, and increased risk of hospitalisation, side effects and adverse drug reactions, which in turn can contribute to increased risk of mortality. 28 , 29 , 30 Polypharmacy represents a public health challenge; as there is an increasing prevalence of long‐term conditions, the burden is expected to increase. 28 Effective management and adjustment of prescriptions is an important consideration to decrease the associated risks and healthcare costs. 28 This study highlights the benefits of weight loss on polypharmacy use in a population with MLTCs and obesity at baseline, thus indicating how treatment of obesity could reduce polypharmacy burden. Similarly, weight loss in this study was associated with fewer clinical consultations compared with individuals with stable weight. These results follow previous findings showing an association between weight loss and reductions in HCRU and the number of prescriptions. 18
However, despite the favourable associations observed between weight loss and the risk of ORCs, polypharmacy and clinical consultations, this study showed an association between weight loss and adverse mental‐health outcomes, hospitalisation and mortality. The higher hazard of hospitalisation and mortality observed in individuals with BMI reductions in this analysis could be related to the method and underlying cause of weight loss. The deleterious effects of unintentional weight loss have been well described, and include increased risk of mortality, as well as increased risk of in‐hospital and specific, disease‐related complications. 31 , 32 , 33 However, intentional weight loss in people with ORCs has been associated with reduced risk of mortality. 33 In people with obesity, weight loss by metabolic surgery is associated with reduced risk of mortality compared with those without surgery. 34 , 35 Similarly, weight loss by pharmacological intervention is associated with reduced mortality in people with overweight/obesity and cardiovascular conditions, compared with no pharmacological treatment. 25 , 36 , 37 Two meta‐analyses describe the possible benefits of lifestyle interventions on all‐cause mortality in people with obesity. 38 , 39 , 40 Reduced incidence of hospitalisations has also been noted with pharmacological intervention‐assisted weight loss, although not with metabolic surgery. 25 , 34
The relationship between weight loss and mental health is of particular interest. Although weight loss by lifestyle intervention and pharmacotherapy has generally been linked to improvements in mental‐health outcomes or quality of life, 41 , 42 , 43 , 44 , 45 the picture for metabolic surgery is mixed, and surgery has been associated with increased risk of adverse mental‐health outcomes. 46 Given the high burden of mental‐health disorders in people with obesity, 47 there is a need to be vigilant and provide continuous monitoring when people with obesity undergo weight loss interventions, especially when larger magnitudes of weight loss are being targeted. It is therefore important that these individuals are offered psychological support when needed, similar to the support given pre‐ and post‐metabolic surgery.
In this study, the association between weight loss and some adverse outcomes was likely due to underlying disease severity driving unintentional weight loss. When examining weight loss as an exposure, it must be ensured that it was intentional, stemming from interventions such as diet, exercise or medication, and not unintentional, caused by other factors such as cancer. Due to the observational nature of our study, this could not be completely controlled. We did not exclude participants taking weight loss medications, anticoagulants or antihypertensives from our analyses in order to retain a population most likely to be proactive in their weight loss efforts and ensure the results were generalisable. However, as these medications can directly influence exposure (i.e., BMI change) and metabolic or cardiovascular outcomes, we conducted a sensitivity analysis in a cohort not taking these medications to address any residual confounding, which supported the primary results. Despite our best efforts to exclude unintentional weight loss, based on specific coding to identify events or conditions that can contribute to weight loss and decreased health status, it is difficult to completely remove these effects in a study based on data collected routinely or for other purposes. However, future studies investigating second‐generation pharmacotherapy for larger magnitudes of weight loss will provide further information about the risk–benefit ratio of these magnitudes of weight loss on health outcomes and HCRU.
Using CPRD Aurum allowed evaluation of data from over 60 million UK patients, and provided a representative sample, given it has been recorded that 98% of the UK population is registered with a GP. 48 However, several limitations are inherent in database studies. One such limitation is that not all information is collected in the database: we were able to collect coded data but could not extract free‐text data. However, in this study, information on time of onset, duration of obesity or, for most individuals, weight loss method was available. Longer duration of obesity is associated with higher risk for multiple, complex ORCs. 49 We were also missing information on the severity of each specific ORC; thus, it was assumed that severity was equivalent between all conditions. In addition, it was challenging to capture alcohol consumption accurately without a significant amount of missing or inaccurate data. Therefore, we decided to only focus on the precise clinical diagnosis of ‘alcoholism,’ which represents the highest‐risk subgroup within the ‘alcohol consumption’ population. However, restricting the definition to only include those with an alcohol‐related diagnosis may underestimate the potential impact of moderate‐to‐heavy alcohol use in the broader population.
In a population with high disease burden, weight loss can reduce the hazard of developing an additional ORC and can result in a lower number of ORCs, a reduced number of clinical consultations and lower polypharmacy rates. On the other hand, we found U‐shaped associations between weight change and mortality, mental‐health conditions and hospitalisation. This highlights the importance of the choice of intervention for people with obesity and MLTCs, considering outcomes beyond weight loss to maximise the benefits and reduce the risks. Appropriate support and follow‐up should be integrated into weight change interventions.
AUTHOR CONTRIBUTIONS
All authors had full access to all data in the study and had final responsibility for the decision to submit for publication. All authors contributed equally to the data interpretation and manuscript writing, approved the final version of the manuscript, and vouch for data accuracy and fidelity to the protocol.
FUNDING INFORMATION
Novo Nordisk A/S was the study sponsor and was responsible for the study design, preparing the protocol and performing the statistical analyses. This article was drafted under the guidance of the authors, who were responsible for all decisions regarding publication, with medical writing and editorial support paid for by the funder.
CONFLICT OF INTEREST STATEMENT
Kamlesh Khunti is supported by the National Institute for Health Research (NIHR) Applied Research Collaboration East Midlands (ARC EM), the NIHR Global Research Centre for Multiple Long‐Term Conditions, the Multiple Long‐Term Conditions Cross‐NIHR Collaboration, the NIHR Leicester Biomedical Research Centre (BRC) and the British Heart Foundation (BHF) Centre of Excellence. Kamlesh Khunti has also acted as a consultant or speaker for, or received grants for investigator‐initiated studies from Applied Therapeutics, AstraZeneca, Bayer, Boehringer Ingelheim, Daiichi Sankyo, Embecta, Lilly, Merck Sharp & Dohme, Nestlé Health Science, Novo Nordisk, Oramed Pharmaceuticals, Pfizer, Roche, Sanofi‐Aventis and Servier. Alex Mourer, Camilla S. Morgen, Lua Wilkinson, Silvia Capucci and Klaus Kaae Andersen are employees of Novo Nordisk and may hold stock options with Novo Nordisk. Abd A. Tahrani was an employee and shareholder of Novo Nordisk at the time of writing the manuscript. Abd A. Tahrani is currently an employee of Amgen Research Copenhagen (ARC). ARC had no role in this project and manuscript.
Supporting information
Data S1. Supporting Information.
Appendix S2. Supporting Information.
ACKNOWLEDGEMENTS
This study was funded by Novo Nordisk A/S. Medical writing support was provided by Deja Scott‐Shemon, MPH, of Apollo, OPEN Health Communications, and funded by Novo Nordisk A/S, in accordance with Good Publication Practice (GPP) guidelines (www.ismpp.org/gpp-2022). The authors thank Mattis Flyvholm Ranthe, Novo Nordisk A/S, for reviewing the manuscript content.
Khunti K, Mourer A, Capucci S, et al. Change in body mass index and disease burden among people with obesity and multiple long‐term conditions. Diabetes Obes Metab. 2026;28(4):2722‐2734. doi: 10.1111/dom.70446
Abd A. Tahrani and Camilla S. Morgen should be considered joint last authors.
[Correction added on 28 January 2026, after the first online publication: The qualifications for Klaus K. Andersen and Abd A. Tahrani have been corrected in this version.]
DATA AVAILABILITY STATEMENT
This study is based on data from CPRD Aurum obtained under licence from the UK Medicines and Healthcare products Regulatory Agency. Patient data are protected and collected by NHS England as part of their care and support. The interpretations and conclusions contained in this study are those of the authors alone. Electronic health records are classified as ‘sensitive data’ by the UK's Data Protection Act 2018; therefore, information‐governance restrictions prevent data sharing via public deposition. Information about access to CPRD data is available here: https://www.cprd.com/research-applications (accessed 29 July 2024).
REFERENCES
- 1. Pearson‐Stuttard J, Ezzati M, Gregg EW. Multimorbidity—a defining challenge for health systems. Lancet Public Health. 2019;4:e599‐e600. [DOI] [PubMed] [Google Scholar]
- 2. Khunti K, Sathanapally H, Mountain P. Multiple long term conditions, multimorbidity, and co‐morbidities: we should reconsider the terminology we use. BMJ. 2023;383:2327. [DOI] [PubMed] [Google Scholar]
- 3. Vogeli C, Shields AE, Lee TA, et al. Multiple chronic conditions: prevalence, health consequences, and implications for quality, care management, and costs. J Gen Intern Med. 2007;22:391‐395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Redelmeier DA, Tan SH, Booth GL. The treatment of unrelated disorders in patients with chronic medical diseases. N Engl J Med. 1998;338:1516‐1520. [DOI] [PubMed] [Google Scholar]
- 5. Romain AJ, Marleau J, Baillot A. Association between physical multimorbidity, body mass index and mental health/disorders in a representative sample of people with obesity. J Epidemiol Community Health. 2019;73:874‐880. [DOI] [PubMed] [Google Scholar]
- 6. Zhao Y, Wang J, Zhu X, et al. Multimorbidity and polypharmacy in hospitalized older patients: a cross‐sectional study. BMC Geriatr. 2023;23:423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Rodrigues LP, de Oliveira Rezende AT, Delpino FM, et al. Association between multimorbidity and hospitalization in older adults: systematic review and meta‐analysis. Age Ageing. 2022;51:afac155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Agborsangaya CB, Ngwakongnwi E, Lahtinen M, Cooke T, Johnson JA. Multimorbidity prevalence in the general population: the role of obesity in chronic disease clustering. BMC Public Health. 2013;13:1161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Chowdhury SR, Chandra Das D, Sunna TC, Beyene J, Hossain A. Global and regional prevalence of multimorbidity in the adult population in community settings: a systematic review and meta‐analysis. EClinicalMedicine. 2023;57:101860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Valabhji J, Barron E, Pratt A, et al. Prevalence of multiple long‐term conditions (multimorbidity) in England: a whole population study of over 60 million people. J R Soc Med. 2024;117:104‐117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Kivimäki M, Kuosma E, Ferrie JE, et al. Overweight, obesity, and risk of cardiometabolic multimorbidity: pooled analysis of individual‐level data for 120 813 adults from 16 cohort studies from the USA and Europe. Lancet Public Health. 2017;2:e277‐e285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. National Institute for Health and Care Research . Multiple long‐term conditions (multimorbidity): making sense of the evidence. 2021. Accessed September 2, 2025. https://evidence.nihr.ac.uk/collection/making‐sense‐of‐the‐evidence‐multiple‐long‐term‐conditions‐multimorbidity/
- 13. Delpino FM, Dos Santos Rodrigues AP, Petarli GB, et al. Overweight, obesity and risk of multimorbidity: a systematic review and meta‐analysis of longitudinal studies. Obes Rev. 2023;24:e13562. [DOI] [PubMed] [Google Scholar]
- 14. Kivimäki M, Strandberg T, Pentti J, et al. Body‐mass index and risk of obesity‐related complex multimorbidity: an observational multicohort study. Lancet Diabetes Endocrinol. 2022;10:253‐263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Agborsangaya CB, Majumdar SR, Sharma AM, Gregg EW, Padwal RS. Multimorbidity in a prospective cohort: prevalence and associations with weight loss and health status in severely obese patients. Obesity (Silver Spring). 2015;23:707‐712. [DOI] [PubMed] [Google Scholar]
- 16. Haase CL, Lopes S, Olsen AH, Satylganova A, Schnecke V, McEwan P. Weight loss and risk reduction of obesity‐related outcomes in 0.5 million people: evidence from a UK primary care database. Int J Obes (Lond). 2021;45:1249‐1258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Mariam A, Miller‐Atkins G, Pantalone KM, et al. Associations of weight loss with obesity‐related comorbidities in a large integrated health system. Diabetes Obes Metab. 2021;23:2804‐2813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Bojke C, Capucci S, Haase CL, et al. Association between weight loss and health care resource utilization in adults living with obesity: evidence from a UK primary care database. Diabetes Obes Metab. 2023;25:3611‐3620. [DOI] [PubMed] [Google Scholar]
- 19. CPRD . Clinical Practice Research Datalink. 2025. Accessed September 2, 2025. https://www.cprd.com/
- 20. Ho ISS, Azcoaga‐Lorenzo A, Akbari A, et al. Measuring multimorbidity in research: Delphi consensus study. BMJ Med. 2022;1:e000247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. Springer; 2000. [Google Scholar]
- 22. Holst KK, Scheike TH, Hjelmborg JB. The liability threshold model for censored twin data. Comput Stat Data Anal. 2016;93:324‐335. [Google Scholar]
- 23. Khunti K, Schnecke V, Haase CL, et al. Weight change and risk of obesity‐related complications: a retrospective population‐based cohort study of a UK primary care database. Diabetes Obes Metab. 2023;25:2669‐2679. [DOI] [PubMed] [Google Scholar]
- 24. Wing RR, Look Ahead Research Group . Does lifestyle intervention improve health of adults with overweight/obesity and type 2 diabetes? Findings from the Look AHEAD randomized trial. Obesity (Silver Spring). 2021;29:1246‐1258. [DOI] [PubMed] [Google Scholar]
- 25. Lincoff AM, Brown‐Frandsen K, Colhoun HM, et al. Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med. 2023;389:2221‐2232. [DOI] [PubMed] [Google Scholar]
- 26. Kosiborod MN, Petrie MC, Borlaug BA, et al. Semaglutide in patients with obesity‐related heart failure and type 2 diabetes. N Engl J Med. 2024;390:1394‐1407. [DOI] [PubMed] [Google Scholar]
- 27. Gulinac M, Miteva DG, Peshevska‐Sekulovska M, et al. Long‐term effectiveness, outcomes and complications of bariatric surgery. World J Clin Cases. 2023;11:4504‐4512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Khezrian M, McNeil CJ, Murray AD, Myint PK. An overview of prevalence, determinants and health outcomes of polypharmacy. Ther Adv Drug Saf. 2020;11:2042098620933741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Condelius A, Edberg AK, Jakobsson U, Hallberg IR. Hospital admissions among people 65+ related to multimorbidity, municipal and outpatient care. Arch Gerontol Geriatr. 2008;46:41‐55. [DOI] [PubMed] [Google Scholar]
- 30. Mangin D, Lamarche L, Agarwal G, et al. Team approach to polypharmacy evaluation and reduction: study protocol for a randomized controlled trial. Trials. 2021;22:746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. McMinn J, Steel C, Bowman A. Investigation and management of unintentional weight loss in older adults. BMJ. 2011;342:d1732. [DOI] [PubMed] [Google Scholar]
- 32. De Stefani FDC, Pietraroia PS, Fernandes‐Silva MM, Faria‐Neto J, Baena CP. Observational evidence for unintentional weight loss in all‐cause mortality and major cardiovascular events: a systematic review and meta‐analysis. Sci Rep. 2018;8:15447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Harrington M, Gibson S, Cottrell RC. A review and meta‐analysis of the effect of weight loss on all‐cause mortality risk. Nutr Res Rev. 2009;22:93‐108. [DOI] [PubMed] [Google Scholar]
- 34. Wiebe N, Tonelli M. Long‐term clinical outcomes of bariatric surgery in adults with severe obesity: a population‐based retrospective cohort study. PLoS One. 2024;19:e0298402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Zhou X, Yu J, Li L, et al. Effects of bariatric surgery on mortality, cardiovascular events, and cancer outcomes in obese patients: systematic review and meta‐analysis. Obes Surg. 2016;26:2590‐2601. [DOI] [PubMed] [Google Scholar]
- 36. Scirica BM, Lincoff AM, Lingvay I, et al. The effect of semaglutide on mortality and COVID‐19‐related deaths: an analysis from the SELECT trial. J Am Coll Cardiol. 2024;84:1632‐1642. [DOI] [PubMed] [Google Scholar]
- 37. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387:205‐216. [DOI] [PubMed] [Google Scholar]
- 38. Kritchevsky SB, Beavers KM, Miller ME, et al. Intentional weight loss and all‐cause mortality: a meta‐analysis of randomized clinical trials. PLoS One. 2015;10:e0121993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Ma C, Avenell A, Bolland M, et al. Effects of weight loss interventions for adults who are obese on mortality, cardiovascular disease, and cancer: systematic review and meta‐analysis. BMJ. 2017;359:j4849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Look AHEAD Research Group . Effects of intensive lifestyle intervention on all‐cause mortality in older adults with type 2 diabetes and overweight/obesity: results from the Look AHEAD study. Diabetes Care. 2022;45:1252‐1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Jones RA, Lawlor ER, Birch JM, et al. The impact of adult behavioural weight management interventions on mental health: a systematic review and meta‐analysis. Obes Rev. 2021;22:e13150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Peckmezian T, Hay P. A systematic review and narrative synthesis of interventions for uncomplicated obesity: weight loss, well‐being and impact on eating disorders. J Eat Disord. 2017;5:15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Fabricatore AN, Wadden TA, Higginbotham AJ, et al. Intentional weight loss and changes in symptoms of depression: a systematic review and meta‐analysis. Int J Obes (Lond). 2011;35:1363‐1376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Cooper DH, Ramachandra R, Ceban F, et al. Glucagon‐like peptide 1 (GLP‐1) receptor agonists as a protective factor for incident depression in patients with diabetes mellitus: a systematic review. J Psychiatr Res. 2023;164:80‐89. [DOI] [PubMed] [Google Scholar]
- 45. Rubino D, Bjorner JB, Rathor N, et al. Effect of semaglutide 2.4 mg on physical functioning and weight‐ and health‐related quality of life in adults with overweight or obesity: patient‐reported outcomes from the STEP 1–4 trials. Diabetes Obes Metab. 2024;26:2945‐2955. [DOI] [PubMed] [Google Scholar]
- 46. Law S, Dong S, Zhou F, Zheng D, Wang C, Dong Z. Bariatric surgery and mental health outcomes: an umbrella review. Front Endocrinol (Lausanne). 2023;14:1283621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Jaison KI, Asharaf H, Thimothy G, et al. Psychological impact of obesity: a comprehensive analysis of health‐related quality of life and weight‐related symptoms. Obes Med. 2024;45:100530. [Google Scholar]
- 48. NHS England . GP appointments data dashboard. 2023. Accessed September 2, 2025. https://digital.nhs.uk/dashboards/gp-appointments-data-dashboard
- 49. Gong HJ, Tang X, Zhou JB. The association between weight change patterns and obesity‐related complex multimorbidity: evidence from NHANES. Front Endocrinol (Lausanne). 2024;15:1400204. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1. Supporting Information.
Appendix S2. Supporting Information.
Data Availability Statement
This study is based on data from CPRD Aurum obtained under licence from the UK Medicines and Healthcare products Regulatory Agency. Patient data are protected and collected by NHS England as part of their care and support. The interpretations and conclusions contained in this study are those of the authors alone. Electronic health records are classified as ‘sensitive data’ by the UK's Data Protection Act 2018; therefore, information‐governance restrictions prevent data sharing via public deposition. Information about access to CPRD data is available here: https://www.cprd.com/research-applications (accessed 29 July 2024).