A favorable effect of hydroxychloroquine on glucose and lipid metabolism beyond its anti-inflammatory role

Abstract

Hydroxychloroquine (HCQ), a commonly used antimalarial drug in rheumatic diseases, has shown favorable metabolic effects on both glucose control and lipid profiles. We describe a case of a young woman with type 1 diabetes whose glycemic control was optimized with the introduction of HCQ as a treatment for her Sjogren syndrome in addition to a subtle yet measurable improvement in her lipid profile. An increasing body of evidence supports the beneficial impacts of HCQ in various ancillary conditions, including diabetes mellitus and dyslipidemia. However, mechanisms of action responsible for these effects remain ill-defined and may include alterations in insulin metabolism and signaling through cellular receptors. These favorable metabolic effects of HCQ and further understanding of underlying mechanisms may provide an additional rational for its use in rheumatic diseases, conditions associated with an elevated cardiovascular risk.

Introduction

Antimalarials such as hydroxychloroquine (HCQ), are among the oldest prescribed drugs still used in clinical practice. Relatively inexpensive and well tolerated, these drugs have been recognized to be effective in autoimmune diseases such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). Interestingly, there is growing evidence of their beneficial impact on cardiovascular risk, particularly diabetes and dyslipidemia. Herein, we present a case on the possible beneficial impact of HCQ on glycemic control in diabetes mellitus and possibly lipid parameters as a trigger for this extensive review of the available literature.

Case presentation

A 24-year-old woman diagnosed at the age of 11 years with type 1 diabetes mellitus presented at the age of 15 years in 2003 to our institution for continuity of her diabetes care. Her glycemic control was suboptimal despite an adjustment of her insulin doses. In September 2011, she was diagnosed with Sjogren syndrome and was started on HCQ 200 mg once daily. An improvement in her glycemic control as evidenced by self glucose monitoring was noted within 1 month of HCQ therapy, with no remarkable hypoglycemic episodes. Her glycated hemoglobin A1C (HbA1C) was successfully reduced to target in January 2011. No C-peptide levels were obtained before and after HCQ treatment since the patient has longstanding type 1 diabetes and has no insulin reserve left. Interestingly, a slight improvement in her low-density lipoprotein cholesterol (LDL-C) was also seen a few months after HCQ initiation. After reviewing the literature, our case seems to be the first case report of a patient with type 1 diabetes who had improvement in her glycemic control while on HCQ therapy.

Literature search

A comprehensive electronic literature search was undertaken of the following medical-related databases: PubMed and Google scholars. The search was conducted for relevant English language publications until May 2014 using combinations of the following search terms ‘antimalarials’, ‘chloroquine’, ‘diabetes’, ‘dyslipidemia’, ‘hydroxychloroquine’, ‘anti-inflammatory treatment’, ‘cardiovascular diseases’ and ‘hypoglycemia’. References from the articles identified by this search strategy and publications available in the authors’ libraries were also used. In addition, the clinical trials website was included in order to capture any additional studies that could address this topic. We undertook supplementary searches in the previously mentioned databases including only one keyword per row to ensure that no literature was missed. All the abstracts were read; for the review, full papers were obtained.

Antimalarials and effects on glucose metabolism

The antidiabetic effect of chloroquine (CQ) was first described in 1984 in a patient with severe insulin resistance in whom the addition of CQ dramatically reduced his insulin requirements [Blazar et al. 1984]. Later on, Smith and colleagues reported on six patients with mild noninsulin-dependent diabetes treated with a short course of CQ with significant improvement in their glucose tolerance [Smith et al. 1987]. Furthermore, the addition of HCQ to either insulin or glibenclamide in the treatment of refractory noninsulin-dependent diabetes for 6 months resulted in a significant decrease in HbA1C by 3.3% compared with placebo as well as a reduction in insulin dose by 30% in the insulin-treated group [Quatraro et al. 1990]. Similarly, patients with sulfonylurea-refractory type 2 diabetes mellitus (T2DM), treated with HCQ, demonstrated an improvement in their glycemic control during the first 6 months of therapy with a reduction of 1.02% in their HbA1C compared with placebo [95% confidence interval (CI) 0.24–1.81] [Gerstein et al. 2002]. A similar but lower decrease in HbA1C of 0.66% was observed in 45 patients with diabetic RA treated with HCQ within 12 months post treatment (95% CI 0.26–1.05) [Rekedal et al. 2010]. Severe hypoglycemia has been reported in patient with insulin dependence and T2DM within 2 weeks of starting HCQ at a dose of 400 mg daily for his polyarthritis. The patient had a subsequent decrease in his insulin requirements by 37% [Shojania et al. 1999]. Another report emphasizing the glucose-lowering effect of HCQ is that describing a case of hypoglycemia in a patient with SLE and T2DM after initiation of HCQ at a dose of 200 mg twice daily with eventual discontinuation of her subcutaneous insulin [Kang et al. 2009].

In vitro and animal studies have shown that HCQ and the parent drug CQ affect insulin metabolism. CQ increases insulin binding to its receptor and alters hepatic insulin metabolism, potentiating insulin action [Pease et al. 1985; Bevan et al. 1995]. Streptozocin-induced diabetic rats treated with HCQ had lower insulin clearance and subsequently lower glucose levels [Emami et al. 1999a, 1999b]. Long-term use of CQ enhanced insulin secretion in rats [Asamoah et al. 1990]. In addition, a 3-day use of CQ at high doses of 250 mg four times daily reduced insulin clearance by 39% and increased secretion of C peptide by 17% in patients with noninsulin-dependent T2DM [Powrie et al. 1991]. This is in contrast to the report by Quatraro and colleagues in which both fasting and glucagon-stimulated C peptide were unaltered after 6 months of treatment with HCQ at a dose of 200 mg three times daily [Quatraro et al. 1990]. Furthermore, in a recent small pilot study on subjects with obesity but without diabetes and systemic inflammatory conditions, treatment with HCQ for 6 weeks was associated with a significant improvement in insulin sensitivity. Markers of inflammation including C-reactive protein and interleukin 6 did not change, arguing in favor of a direct effect of HCQ on insulin metabolism rather than a consequence of a reduction in inflammation [Mercer et al. 2012]. The anti-inflammatory mechanism of action of antimalarials in the treatment of patients with RA is unknown but is thought to involve changes in antigen presentation or effects on the innate immune system. This might also contribute to reduction in inflammation and consequently reducing metabolic parameters such as glucose and possibly lipids, which might be of benefit in reducing cardiovascular risk and mortality [Fox, 1993].

In addition, among patients with SLE, HCQ users had a lower frequency of metabolic syndrome [Bellomio et al. 2009]. Besides its glucose-lowering effect in patients with diabetes, lower fasting blood sugar levels [Penn et al. 2010], as well as hypoglycemia have also been reported with CQ and HCQ in subjects without diabetes [Abu-Shakra and Lee, 1994; Cansu and Korkmaz, 2008; Winter et al. 2011]. Interestingly, HCQ lowers the incidence of developing diabetes mellitus. In a retrospective cohort of 1127 patients with RA, the use of HCQ was associated with a 71% reduction in the risk of incident diabetes [hazard ratio (HR) 0.29; 95% CI 0.09–0.95; p = 0.04] [Bili et al. 2011]. These findings replicated those reported by Wasko and colleagues, who demonstrated a 77% reduced incidence of diabetes in patients with RA who took HCQ for longer than 4 years, after adjustment for risk factors of diabetes, disease activity and glucocorticoid use [Wasko et al. 2007]. Similarly, another recent retrospective cohort of 13,905 patients with a diagnosis of either RA or psoriasis reported a lower incidence of diabetes among HCQ and tumor necrosis factor inhibitor users compared with other nonbiological disease-modifying antirheumatic drugs (HR 0.54; 95% CI 0.36–0.80) [Solomon et al. 2011].

Other anti-inflammatory medications have shown similar effects of HCQ on glucose and lipid metabolism, such as salsalate. Two randomized placebo-controlled studies were conducted to evaluate the effect of salsalate on insulin resistance and cardiovascular risk factors. The first study showed that in people with abnormal glucose tolerance, salsalate therapy was well tolerated, lowered fasting glucose, increased adiponectin and reduced adipose tissue nuclear factor κB. These changes were not related to changes in peripheral insulin sensitivity, suggesting additional mechanisms for metabolic improvement [Goldfine et al. 2013a]. A larger multicenter, placebo-controlled trial conducted on 286 patients showed that salsalate improves glycemia in patients with T2DM and decreases inflammatory mediators [Goldfine et al. 2013b].

The mechanism by which HCQ works to improve glycemic control in patients with type 1 diabetes is not known and our case seems to be the first case report in the literature. Since HCQ is an immunomodulatory drug (has an inhibitory and an immunomodulatory effect on T cells and interleukin 1 and interleukin 6) and type 1 diabetes is an autoimmune disease, a possible reduction in islet-cell autoimmunity by HCQ could be the underlying mechanism since our patient has longstanding type 1 diabetes with no insulin reserve and an effect on insulin resistance is less likely to be the underlying mechanism that led to the improvement in glycemic control [Ben-Zvi et al. 2011].

All of the studies in Table 1 suggest a glucose-lowering effect of antimalarials, the mechanism behind which remains to be fully elucidated.

Table 1.

Effects of antimalarial agents on glycemic control.

Author [year]	Study design	N	Mean Age	BMI or weight	Dose	Duration of therapy or time of F/U	Findings
Quatraro [1990]	Prospective, randomized, placebo controlled	38 T2DM 22 on insulin 16 on glibenclamide	58	27.9 ± 3.2 kg/m2	200 mg three times a day	6 months	HCQ reduced insulin dose by 30% and HbA1C ↓ 3.3%
Powrie [1991]	Prospective, randomized, placebo controlled	20 T2DM on diet 10 CQ 10 placebo	54	26 kg/m2	250 mg four times a day	3 days	↓ mean plasma glucose by 35 mg/dl (p < 0.01) in CQ users ↓ insulin clearance rates by 39% (p < 0.01) in CQ users
Gerstein [2002]	Randomized, placebo controlled	135 T2DM 69 HCQ	57.5	31.9 kg/m2	Up to 300 mg twice a day	6 months	↓ 1.02% in HbA1C versus placebo (95% CI 0.24–1.81) LDL-C ↓ more in HCQ users p < 0.0001 TC ↓ more in HCQ users p = 0.005 TG ↓ more in HCQ users p < 0.0001 HDL-C NS
Wasko [2007]	Prospective observational for 21.5 years	4905 RA 1808 HCQ users 3907 non HCQ	53.7 58.2	28.0 kg/m2	340 mg once per day	3.1 years	HR of 0.62 (95% CI 0.42–0.92) in HCQ users compared with nonusers
Rekedal [2010]	Observational cohort	85 RA + T2DM 45 HCQ	61	35.4 ± 8.5 kg/m2	NA	3–12 months	↓ HbA1C of 0.66% compared with pretreatment (95% CI 0.26–1.05)
Penn [2010]	Cross-sectional	Nondiabetic 149 SLE 48% HCQ 177 RA 18% HCQ	49.8 ± 9.7	27.6 ± 6.2 kg/m2	200–400 mg daily	NAMean disease duration = 16 years	Lower serum glucose in SLE HCQ users (p = 0.04) and RA HCQ users (p = 0.05) versus nonusers
			58.5 ± 10.4	27.8 ± 5.8 kg/m2			Lower HOMA-IR in SLE (p < 0.05)
Solomon [2011]	Retrospective cohort	13,905 RAor psoriasis 5682 HCQ 8195 MTX	61	NA	NA	Mean F/U of 5.8 months	Lower incidence of diabetes among HCQ users; HR 0.54 (95% CI 0.36–0.80) compared with other nonbiological DMARDs
Bili [2011]	Retrospective cohort	1127 RA 333 HCQ	60.7	29.1 kg/m2	400 mg once per day	Mean duration of HCQ exposure is 14 months	Lower risk of DM among HCQ users HR 0.29 (95% CI 0.09–0.95, p = 0.04)

BMI, body mass index; CI, confidence interval; CQ, chloroquine; DMARD, disease-modifying antirheumatic drug; F/U, follow up; HbA1C, glycated hemoglobin A1C; HCQ, hydroxychloroquine; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, HOMA insulin resistance; HR, hazard ratio; LDL-C, low-density lipoprotein cholesterol; MTX, methotrexate; NA, not available; NS, not significant; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; TC, total cholesterol; TG, triglyceride; T2DM, type 2 diabetes mellitus.

Antimalarials and effects on lipid metabolism

There is increasing evidence demonstrating the beneficial impact of antimalarial agents on lipid metabolism besides glycemic control (Table 2). In rat hepatocytes, CQ was shown to be an effective inhibitor of cholesterol synthesis [Beynen et al. 1981]. CQ upregulated LDL-C receptors, enhancing the plasma removal of this lipoprotein and resulting in lowering of its serum levels [Sachet et al. 2007]. In fact, the influence of CQs on lipid metabolism has been evaluated in several retrospective as well as prospective studies [Kavanaugh et al. 1997; Munro et al. 1997]. Most of the studies have reported a favorable effect of antimalarials on serum lipid levels manifested by a reduction in LDL-C and total cholesterol (TC) and an increase in high-density lipoprotein cholesterol (HDL-C), notably in those treated with steroids [Wallace et al. 1990; Hodis et al. 1993; Petri et al. 1994; Rahman et al. 1999; Tam et al. 2000; Borba and Bonfa, 2001]. Interestingly, changes in lipoproteins can be detected as early as 3 months after antimalarial therapy [Rahman et al. 1999; Cairoli et al. 2012]. Reductions up to 54% in triglycerides (TGs), TC and apolipoprotein CIII levels have been reported in HCQ users in 18 women with mild or inactive SLE compared with nonusers. Furthermore, in a large longitudinal study involving 1260 patients with SLE, the use of antimalarials was negatively correlated with TC (p < 0.0001) [Nikpour et al. 2010]. Similarly, the use of HCQ in patients with RA was associated with lower TC, LDL-C, TC/HDL-C and LDL-C/HDL-C ratios [Toms et al. 2011]. Consistent with these results, another recent study that analyzed a cohort of 706 patients with RA of a median duration of 1.98 years found a significant decrease in LDL-C, TC, LDL-C/HDL-C and TC/HDL-C with HCQ use [Morris et al. 2011]. In addition, in a prospective randomized trial, Mundor and colleagues reported an overall increase of 15% in HDL-C in HCQ users in a population of patients with RA after 12 months of therapy compared with a decrement of 12% in patients treated with gold (p = 0.006) [Munro et al. 1997]. Reduced levels of apolipoprotein B lipoproteins were also observed in patients with RA and SLE treated with CQ [Vazquez-Del Mercado et al. 2002] and compared with untreated patients and healthy subjects [Munoz-Valle et al. 2003]. Therefore, antimalarials could reduce serum lipid levels and interestingly diastolic blood pressure [Rho et al. 2009] and possibly reduce atherosclerosis and coronary artery disease risk associated with rheumatic diseases [Petri et al. 1994]. However, in a recent systemic review of nine studies, the majority of which were cross sectional, the authors reported a weak effect of antimalarials on lipid profile in patients with SLE. Furthermore, two studies from China and Iran failed to show a positive effect of HCQ on lipid profile [Tam et al. 2000; Karimifar et al. 2007]. Similarly, in a recent study by Rossoni and colleagues evaluating the effect of CQ on cholesterol levels in a Brazilian SLE population, no significant differences were observed in CQ users compared with nonusers after adjustment for statin and corticosteroid use [Rossoni et al. 2011]. The observed inconsistencies could be related to the cross-sectional nature of the studies, the heterogeneity of the population in terms of disease activity, duration and ethnicity, lack of adjustment in some of confounding factors such as steroids and use of statin.

Table 2.

Effects of antimalarial agents on lipid profile.

Author [year]	Study design	N	Mean age	BMI or weight	Dosing	Duration of therapy or time of F/U	Findings
Hodis [1993]	Cross sectional	18 SLE 9 HCQ users	30	59.24 ± 2.7kg	200–400 mg once per day	4–10 years mean = 43 months	Lower TG (p < 0.001), VLDL-C (p < 0.001), apoCIII (p < 0.01) in HCQ group
Petri [1994]	Longitudinal cohort	264 SLE 47.2% HCQ users	38.8 ± 12.2	163 ± 40 lb	200–400 mg once per day	NA	Lower serum cholesterol level by 8.94% in HCQ users
Kavanaugh [1997]	Double blind, prospective	17 SLE five placebo six 400 mg HCQ six 800 mg HCQ	NA	NA	400–800 mg	NA	↓ in TC in 400 mg group by 11.6 mg/dl (p = 0.03) ↓ in TC by 13.4 mg/dl, TG, VLDL-C in the 800 mg group (p = 0.05)
Munro [1997]	Prospective, randomized	100 RA 51 HCQ	51	64.6 kg	400 mg for 6 months then 200 mg for another 6 months	3, 6, 12 months	↑ in HDL-C by 15% in HCQ users (p = 0.04)
Rahman [1999]	Prospective cohort Cross sectional	53 AM 29 AM + PRED 38 AM + PRED 36 AM 201 PRED 181 PRED + AM	NA	NA	Mean CQ dose = 4 mg/kg/day (max 250 mg/day) Mean HCQ dose = 6 mg/kg/day (max = 400 mg/day)	3, 6 months	↓ in TC by 4.1% at 3 months in AM users (p = 0.02). At 6 months ↓ TC NS in AM users ↓ TC 11.3% at 3 months (p = 0.0002) and 9.4% at 6 months (p = 0.004) in AM + PRED users Change in TC lower in AM users who initiated PRED compared with non AM users (18% versus 32%, p = 0.0149) Lower TC by 11% in AM + PRED users compared with PRED users (p = 0.0023)
Tam [2000]	Cross sectional	123 SLE 34 HCQ 25 CQ	45.3	68.6 ± 15.9 kg	Mean HCQ dose =355 ± 84 Mean CQ dose = 205 ± 56	Mean HCQ use 38 ± 58 months Mean CQ use 123 ± 83 months	Lower TC (p = 0.002), VLDL-C (p = 0.01), LDL-C (p = 0.007) in AM users
Tam [2000]	Cross sectional	65 SLE 44 HCQ	39 ± 8	21.7 ± 3.4 kg/m2	Mean dose = 244 ± 86	NA	No significant difference between HCQ users and nonusers
Borba [2001]	Cross sectional	60 SLEtherapy 17 no 14 CQ 15 PRED 14 CQ + PRED 30 controls	NA	33.7 ± 7.3 kg/m2 35.2 ± 5.4 kg/m2 34.6 ± 6.94 kg/m2 32.2 ± 10.7 kg/m2 31.2 ± 5.5 kg/m2	250 mg daily	>1 year	Higher HDL-C in CQ users compared with no therapy (p < 0.05) but similar to healthy controls Higher HDL-C and lower TG and VLDL-C in PRED + CQ compared with CQ p < 0.05
Munoz-Valle [2003]	Cross sectional	61 RA 57 SLE 59 CQ 50 healthy controls	38 32	NA	150 mg once a day	NA	Lower apoB levels (–33 mg/dl, p < 0.05) in CQ users
Wallace [1990]	Retrospective	155 SLE or RA 58 HCQ 35 steroids 18 HCQ + steroids 44 no therapy	46.6	63.3 kg	Mean dose = 386 mg	NA	Lower LDL-C (p = 0.0039) and TC (p = 0.007) in the HCQ group compared with no therapy 15% lower LDL-C (p < 0.05) and TG (p < 0.05) in HCQ + steroid group compared with steroid group alone
Sachet [2007]	Cross sectional, controlled	20 SLE 10 CQ 10 no therapy 10 healthy	35.4 ± 7.5 36.5 ± 6.9 35.6 ± 8.9	23.0 ± 2.5 kg/m2 24.3 ± 2.4 kg/m2 22.2 ± 3.7 kg/m2	250 mg daily	7.1 ± 2.9 months	Lower TC, LDL-C, HDL-C in CQ group compared with no therapy group and controls (p < 0.05) Higher HDL-C compared with no therapy group (p < 0.05)
Karimifar [2007]	Cross sectional	41 HCQ 15 CQ	29	24 kg/m2	NA	NA	No significant difference in TG between HCQ users and nonusers Lower TG in CQ users compared with nonusers HDL-C NS
Rho [2009]	Cross sectional	169 RA 42 HCQ	54.2 ± 11.8	NA	NA	NA	LDL-C (p = 0.03), TG (p = 0.03), diastolic BP (p = 0.02) lower in HCQ users
Morris [2011]	Cross sectional	706 256 HCQ	Median age of 65	Median BMI = 29.8 kg/m2	200–400 mg/day	Median exposure time to HCQ 1.98 years	HCQ users: LDL-C ↓ 7.55 mg/dl (p < 0.001) HDL-C ↑1.02 mg/dl (p = 0.2) TC ↓7.7 mg/dl (p = 0.002) TG↓10.91 mg/dl (p = 0.06) LDL-C/HDL-C ↓ 0.136 (p = 0.008) TC/HDL-C ↓ of 0.191 (p = 0.006)
Rossoni [2011]	Cross sectional	60 SLE 34 CQ or HCQ	48.7	25.5 kg/m2 (women) 24.5 kg/m2 (men)	400 mg HCQ 250 mg CQ	At least 6 months	No significant difference in cholesterol and HDL-C levels between CQ users and nonusers
Cairoli [2012]	Longitudinal	24 SLE HCQ	37.2 ± 16	27.1 ± 5.8 kg/m2	200 mg once a day	3 months	TC ↓ 7.6% (p = 0.005); LDL-C ↓ 13.7% (p= 0.036)

AM, antimalarials; apoB, apolipoprotein B; apolipoprotein CIII; BMI, body mass index; BP, blood pressure; CQ, chloroquine; F/U, follow up; HCQ, hydroxychloroquine; HDL-C, high-density lipoprotein cholesterol; HOMA-IR, HOMA insulin resistance; HR, hazard ratio; LDL-C, low-density lipoprotein cholesterol; MTX, methotrexate; NA, not available; NS, not significant; PRED, prednisone; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; TC, total cholesterol; TG, triglyceride; VLDL, very low density lipoprotein cholesterol.

The effects of HCQ on atherosclerosis (AS) and vascular disease in chronic kidney disease (CKD) are not known yet. A study by Guin and colleagues showed that subclinical AS and endothelial dysfunction are demonstrable features even in early RA which improved after therapy. Early intervention of RA with Disease-Modifying Antirheumatic Drugs (DMARDs) not only controls the disease but also retards the atherosclerotic progression [Guin et al. 2013]. A trial to prove that HCQ treatment in individuals with CKD will provide clinically significant benefit in the management of cardiovascular disease and will provide biological and functional atherosclerotic benefits is being conducted on 62 patients. This ‘proof-of-concept’ randomized double-blinded placebo-controlled study will evaluate the nature and extent of HCQ effects and if HCQ is found to be significantly beneficial [ClinicalTrials.gov identifier: NCT01537315]. Another trial that is expected to be completed in 2015 is being conducted on 30 patients and is addressing the metabolic effects of HCQ on blood glucose, blood pressure and blood cholesterol in T2DM. This trial is offering a unique opportunity to develop a novel approach for lowering blood pressure, lipids (cholesterol and TGs) and glucose in people at high risk and possibly lowering their cardiovascular risk and mortality [ClinicalTrials.gov identifier: NCT02026232].

Further well designed studies are needed to clarify the impact of antimalarials on lipid profile and cardiovascular disease management.

Conclusion

This case and review highlight the need to re-examine HCQ as a potential therapy for T2DM and consider its use especially in patients with rheumatism and diabetes. Furthermore, endocrinologists and rheumatologists should be aware of the potential hypoglycemic effect of antimalarials and the need for close monitoring. The favorable lipid-lowering and antidiabetic properties of HCQ renders this drug an attractive medical option. Given the elevated cardiovascular risk associated with RA and SLE, the addition of HCQ to patients’ usual treatment could counteract the dyslipidemic effect of glucocorticoids, resulting in a potential minimization of atheroma progression and thus possibly lowering mortality due to cardiovascular diseases.

In conclusion, HCQ is a relatively safe and inexpensive medication and has a favorable glucose and lipid lowering effect that provides a rationale for its use in addition to its known benefits in rheumatic diseases. Further studies are needed in patients with type 1 diabetes who receive HCQ treatment for other rheumatologic conditions to clarify the mechanism by which HCQ affects their glycemic control.