Metabolic syndrome, or ‘Syndrome X’, was first identified by endocrinologist Gerald Reaven in the 1980s, and the root cause of this condition is described as ‘insulin resistance’.¹² Dr Stephen Phinney and Dr Jeff Volek explain this further as a condition of ‘carbohydrate intolerance’.¹³ According to recent NHANES data, almost one in three Americans now have metabolic syndrome.¹⁴ Metabolic syndrome is a sign of poor carbohydrate metabolism and hyperinsulinaemia and, if not reversed, portends an inevitable decline in overall health.¹⁵ Markers of metabolic syndrome are abdominal obesity, high blood glucose, high triglycerides (TG), low high-density lipoprotein (HDL) cholesterol and high blood pressure. Other associated conditions are increased inflammation, vascular dysfunction, non-alcoholic fatty liver disease (NAFLD), polycystic ovary syndrome, sleep apnea and some types of cancer and dementia.^{16,17,18,19,20,21,22,23,24} Obesity is commonly viewed as the cause of metabolic syndrome. However, while obesity is highly correlated with the condition, we (and others) suggest that it is the expansion of adipose stores beyond a person’s individual storage capacity,^25,26 driven by and in conjunction with hyperinsulinaemia because of the overconsumption of processed dietary carbohydrates in the setting of a Western diet – so prevalent in modern society – that drives this disease process.^26,27,28 Inflammation, especially the presence of pro-inflammatory macrophages in fat stores, such as the visceral adipose tissue, is another key marker of the developing pathology.^25,29 The inflamed adipose tissue appears to lose its ability to properly respond to insulin, resulting in loss of its capacity to protectively buffer fatty acids (including sugars and starches converted to fatty acids via de novo lipogenesis to be ‘safely’ stored).^25,29 This leads to reduced capacity to clear circulating glucose, as well as greater hepatic exposure to gluconeogenic precursors from rapid fatty acid turnover occurring in the insulin-resistant adipose tissue,³⁰ both resulting in the persistently elevated blood glucose associated with metabolic syndrome.³¹ As a population-level intervention, restriction of carbohydrates, a key contributor to the development of metabolic syndrome (especially when processed), should therefore be considered a primary intervention for those with metabolic disease.

The prospect of reversing progressive T2D is somewhat new to both the public and health care practitioners. Many assume the diagnosis is a purely progressive disease,^32,33 including an inevitable march towards worsening health and organ systems failure.³⁴ In the lexicon of diabetes care, ‘intensive management’ has superseded ‘cure’, ‘remission’ or ‘reversal’. The ADA’s 2017 Standards of Medical Care in Diabetes states that: ‘Diabetes is a complex, chronic illness requiring continuous medical care with multifactorial risk-reduction strategies beyond glycaemic control’.³⁵ In line with this, the ADA’s 2013 Nutrition Therapy Recommendations for the Management of Adults with Diabetes states that T2D is ‘progressive’ in nature and ‘nutrition and physical activity interventions alone (i.e. without pharmacotherapy) are generally not adequately effective in maintaining persistent glycaemic control over time for many individuals’.³⁶

In current treatment approaches, pharmacotherapy typically is the focal point and carbohydrate restriction is de-emphasised. For instance, a recent Kaiser Permanente study involving 120 000 patients concluded that prolonged T2D remission in community settings without bariatric surgery is very rare; it occurred in only 0.007% of the study population. During the 7-year study, more study participants (1.7%) died than the percentage who experienced any level of remission, and diabetes-associated medication use, health care costs and complications increased.³⁷ In addition, seven multinational, multicentre, randomised controlled trials aimed at achieving tight blood glucose control with medications failed to demonstrate the expected reductions in heart disease, the major killer of patients with diabetes, or in overall mortality.^{38,39,40,41,42,43,44} In contrast to these outcomes, strong evidence exists for a promising alternative approach.

The first well-recorded version of a low-carbohydrate diet was described in 1863 by the Englishman William Banting, who restricted starchy and sugary foods to overcome obesity on the advice of his physician.⁴⁵ The popularity of the diet as an effective weight loss regimen continued well into the 20th century, when it was common knowledge that ‘sugar and starches are fattening’. By mid-20th century, studies on the low-carbohydrate diet were a rarity in the medical and nutritional literature because of a growing belief that fat, and particularly saturated fat, not carbohydrate, was the more likely dietary culprit.⁴⁶ Although resistance within the medical community persists,⁴⁷ in recent years dietary fat has slowly been exonerated, in accordance with the best available evidence.^{48,49,50,51,52,53} Studies are also now emerging to show the efficacy, sustainability and other positive effects of low-carbohydrate diets, especially in those with metabolic disease.^{54,55,56,57,58,59,60,61}

The definition of a low-carbohydrate diet varies in the literature, but most in the field agree that anything over 130 g – 150 g per day is not low carbohydrate.⁶² An amount of less than 50 g of carbohydrate per day is considered very low carbohydrate, which would put most adults into nutritional ketosis.⁶³ During this state, the body relies primarily on fatty acids and ketone bodies produced from fat stores, not glucose, for energy.

Numerous randomised, controlled trials have shown that well-formulated low-carbohydrate dietary patterns are highly effective for treating obesity and improving a spectrum of risk factors typical of patients with metabolic syndrome and diabetes who previously ate a traditional diet.^{2,64,65,66,67,68,69,70,71} An example is a report by McKenzie et al. on the interim results of their 2-year outpatient trial on T2D patients adhering to a ketogenic diet (very low carbohydrate, usually < 50 g/day).¹³ Although not a randomised, controlled trial, the trial demonstrates the wide-scale applicability and efficacy of carbohydrate restriction. At 10 weeks, 91% of the 262 patients remained active in the trial and had experienced an average HbA1c reduction from 7.6% to 6.5% and an average of 7.2% body weight loss. More than half of the participants (56.8%) had reduced or eliminated one or more diabetes medications. At 6 months, 89% of the participants were retained, and weight loss averaged 12%. These findings are congruent with the physiology of insulin accelerating intracellular storage of glucose and fats, whereas reduced serum insulin from a low-carbohydrate approach allows lipolysis and gives the body access to its stored fat as an energy source.⁷² In contrast, intensive management of T2D with the ADA-recommended low-fat diet and standard drug regimen routinely results in minimal weight loss even when glucose control is optimised.³⁸

In a critical review of the literature, Feinman et al. present 12 points of evidence (summarised below) to support the low-carbohydrate diet as the best first-line strategy for treating T2D and the most effective adjunct to pharmacology in type 1 diabetes (T1D)⁵⁵:

Despite a growing body of evidence for low-carbohydrate diets’ superior results in the management of obesity, cardiovascular disease risk, metabolic syndrome and T2D, objections by the medical and public health communities persist.⁷³ This is frequently based on flawed or outdated evidence.⁷⁴ For example, misunderstanding about the body’s need for glucose, in particular by the central nervous system, continues despite basic scientific knowledge to the contrary.⁵ In concert, the ADA does not recommend a low-carbohydrate diet for diabetes management. While recognising that ‘carbohydrate intake has a direct effect on postprandial glucose levels’ and ‘total amount of carbohydrate eaten is the primary predictor of glycaemic response’, the ADA advises that ‘a variety of eating patterns (combinations of different foods or food groups) are acceptable for the management of diabetes’ and recommends adjusting carbohydrate intake to mealtime insulin dosing.³⁶ The ADA has refrained from making a robust recommendation for carbohydrate restriction. It posits that ‘evidence is inconclusive for an ideal amount of carbohydrate intake for people with diabetes’.³⁶

Although a reduced-carbohydrate approach may theoretically improve health, if it is too difficult to follow long term or if it brings with it unwanted effects, the idea that people should reduce their carbohydrates is at best irrelevant. Our hypothesis is that if one adheres to a well-formulated low-carbohydrate diet, the success and health benefits can be maintained over years and even for a lifetime. To probe the real-life utility of such an approach, previous examinations of people following some kind of carbohydrate restriction have been conducted. For example, Tim Noakes, a South African physician and low-carbohydrate diet researcher, published an analysis of more than 100 communications he had received from people following some kind of carbohydrate restriction. The participants reported reduced weight, hunger, irritable bowel syndrome symptoms, hypertension and medication needs, as well as improved glycaemic control and exercise capacity.⁷⁵ A survey of more than 2000 members of an online, low-carbohydrate support group found positive health outcomes for participants, such as decreased body weight and an improved lipid profile.⁷⁶ A survey of more than 2500 people following a low-carbohydrate diet of some type (or a paleo diet, which is often a reduced-carbohydrate diet) found a variety of positive health benefits such as reduced weight, hunger, irritable bowel syndrome symptoms, joint pain, arthritis pain, brain fog and acne, as well as improved glycaemic control, psychological well-being, athletic performance and energy.⁷⁷ About 100 people from the National Weight Control Registry, which tracks people who have long-term weight loss success, reported using a low-carbohydrate diet approach and reported long-term weight loss with reduced hunger.⁷⁸

The purpose of this research was to gain a detailed, updated understanding of adults who were already voluntarily following a low-carbohydrate diet. Because long-term dietary trials are complex and costly, we were especially interested in the experience of long-term diet adherents. Such research is hypothesis-generating and enables researchers to explore the actual lived experience of people following such a dietary approach.

The Low Carb Lifestyle Survey was a quality improvement project for the American Board of Family Physicians (ABFM), developed by Dr Mark Cucuzzella and with survey design assistance from collaborators in the United Kingdom (diabetes.co.uk).

The survey was shared with the American and international ‘low-carb’ community via social media and through international health care professionals who support therapeutic uses of low-carbohydrate diets. As Dr Cucuzzella has been recommending a low-carbohydrate diet as an option to his patients with T2D for 6 years, the survey was also shared with individual patients he sees and members of his local community.

The 57-question survey was administered on Survey Monkey (an online survey company); the survey was open from December 2016 to June 2017. The survey included closed-ended, multiple-choice and open-ended questions. Responses to the survey were kept confidential. The full 57-question survey can be viewed in supplemental materials (see Appendix 1).

Where appropriate, we used chi-squared tests to compare proportions and paired t-tests to compare means and standard deviation changes within the same groups. Statistical analyses were performed using IBM SPSS software Version 22 (SPSS Inc., Chicago, IL) and GraphPad Prism version 7 (GraphPad Software, La Jolla, CA).

The survey was completed by 1580 respondents. Respondents were required to answer all questions, with the exception of the open-ended items. The majority of respondents resided in the United States, with the rest residing elsewhere (Table 1). Most of the respondents were female, white, middle-aged and working full time. Of the survey respondents, 97% said that they considered themselves to be adherent to a low-carbohydrate diet. More than half of the respondents said that they had been on a low-carbohydrate diet for at least 1 year. The most commonly reported way respondents had learned about the low-carbohydrate diet was through the Internet. Respondents were allowed to choose more than one reason for starting a low-carbohydrate diet. Three out of four respondents said they started the diet for weight loss, but chronic disease improvement and a desire for more energy were also common responses.

Almost half (49%) of the respondents reported a very restricted average carbohydrate intake of less than 30 g per day. Almost one-third (32%) reported an average carbohydrate intake of 30 g – 50 g per day, 16% reported 50 g – 100 g per day and the remaining 3% reported more than 100 g per day. Average carbohydrate intake was inversely associated with amount of time spent on the diet; that is, of those adherents for less than 6 months, 65% reported an average daily carbohydrate intake of less than 30 g. For participants adherent from 6 months to 2 years, daily consumption of less than 30 g daily still comprised the majority response, but had declined with time. By 2 years, those consuming less than 30 g daily and those consuming 30 g – 50 g daily were even at 38% each. This is consistent with one approach to carbohydrate restriction, starting with a very low-carbohydrate intake and then gradually increasing the carbohydrate level as time goes on.⁷⁹ The amount of weekly physical activity reported was an average of 164 min, with 18% reporting no regular physical activity.

The large majority of respondents reported weight loss and waist circumference reduction with a low-carbohydrate diet (Table 2). Three out of four reported weight loss of 10 pounds or more, with more than one-third of respondents reporting a weight loss of more than 30 pounds. Similarly, 81% respondents reported some reduction in waist size, with 17% reporting a reduction of 5 inches or more. Decreases in weight and waist circumference correlated inversely with daily carbohydrate intake, with a greater proportion of participants (those consuming less than 30 g of carbohydrate per day) reporting more than 3 inches lost from the waist or more than 20 pounds total weight loss (Figure 1).

Among respondents who knew their weight and waist circumference before and after diet initiation, at 6–12 months on the diet (Figure 2), 65% reported a weight loss of 20 pounds or more and 52% reported a waist circumference reduction of 3 inches or more. We found similar results for people who reported following a reduced carbohydrate diet for even longer. For example, for those on the diet 2 years or more, 59% reported a weight loss of 20 pounds or more and 46% reported a waist circumference reduction of 3 inches or more. Providing laboratory values was optional in the survey. Lab test values of respondents who knew their values before diet initiation and after are shown in Tables 3 and 4. Improvements were reported for fasting blood glucose, postprandial blood glucose and triglyceride/HDL (TG/HDL) ratio (Table 3).

We also examined the proportion of participants reporting HbA1c levels in the normal range or in the ranges diagnostic of pre-diabetes or T2D (Table 4). We noticed large improvements. For example, the proportion of participants reporting an HbA1c below 5.5 increased almost fourfold (17% to 65%). Almost half (49%) reported an HbA1c of 6.5 or more before diet initiation. After initiation, fewer than one in 10 (7%) reported an HbA1c that high. Absolute increase in participants reporting a normal HbA1c (< 5.5; 48.7% increase) and absolute decrease in participants reporting a diabetic HbA1c (≥ 6.5; 41.4%) were both greatest in the group eating less than 30 g of carbohydrate per day, with smaller absolute improvements with increasing carbohydrate intake (Table 5).

Respondents reported declines in use of medications (e.g., antidepressant, anti-anxiety, sleep aids, pain relief, anti-inflammatory and phosphodiesterase type 5 inhibitors for erectile dysfunction) after starting the diet (Table 6). The use of all types of medications (except for erectile dysfunction) reduced by more than 50%.

Small minorities of respondents reported current use of medications for diabetes, high blood pressure and hyperlipidaemia, with some reporting having reduced or completely discontinued such medications after initiating the diet (Table 7). While more than half of the respondents said that medication costs were not relevant for them, (e.g. NHS on no medication costs or not on medication) 25% reported a savings, with the average reported savings in medications cost being $288 per month.

Respondents reported improvements in energy level (see Figure 3), ability to perform usual activities (such as housework or leisure activities), pain and emotional health after starting a low-carbohydrate diet (Table 8). For example, before initiation, 59% of respondents reported low energy levels and only 4% reported high energy levels, compared to 3% and 51%, respectively, after initiation. The numbers of respondents reporting moderate to extreme problems with mobility, pain or symptoms of anxiety or depression declined after diet initiation (Table 8). Respondents reported improvement in a variety of aspects of physical and psychological well-being after starting a low-carbohydrate diet. For example, at least eight out of 10 respondents reported improvement in outlook, happiness and self-esteem, and six out of 10 reported improved confidence in controlling blood sugar (Table 9). Similarly, before initiating a low-carbohydrate diet, the great majority of respondents reported negative between-meal experiences, such as intense hunger, tiredness and difficulty concentrating (Table 10). Fewer than one in five respondents reported having such experiences after diet initiation.

Respondents were invited to answer three open-ended questions about their experiences on a low-carbohydrate diet. Common responses are summarised in Table 11. Commonly reported health changes were improvements on a wide spectrum: breathing and allergies, markers of NAFLD, libido and reproductive health, mental clarity and digestion. Commonly reported challenges related to a low-carbohydrate diet included negative reactions from family, friends and health professionals; dining out, special occasions and travelling; flu-like symptoms associated with transitioning to a low-carbohydrate diet; and desires for certain foods. Common reports of the important learning or information gained from being on a low-carbohydrate diet pertained to insulin resistance and metabolic syndrome; the difference between nutritional ketosis and ketoacidosis; and awareness about how different foods affect satiety, weight gain and how one feels.

An important marker of metabolic health and adverse long-term outcomes is a large waist circumference and visceral fat.^81,82 Although we had to rely on participants’ self-report of pounds and inches gained or lost, in light of the fact that sustained weight loss is universally difficult, our dramatic findings deserve consideration. A weight loss of 20 pounds or more was reported by 65.1% of those on the diet for 6–12 months and by 58.8% of those on the diet for 2 years or more (Figure 2).

Although fewer than half (44.1%) of respondents reported that their reason for following a low-carbohydrate diet was to improve a chronic condition, we collected robust data for diabetes and glucose control. Of the entire cohort, 495 knew their HbA1c before diet initiation and 413 knew their value post-diet initiation. By definition, people with T2D have an HbA1c of ≥ 6.5, and pre-diabetes is defined by an HbA1c of 5.7–6.4. Many studies suggest that an HbA1c around 5.0–5.5 is likely to be the ideal.^83,84,85,86 Of those reporting their after-initiation values, the proportion with an HbA1c < 6.5 nearly doubled, and nearly four times as many reported an HbA1c of < 5.5. Most importantly, those with an HbA1c in the diabetic range (≥ 6.5) decreased from 48.5% to 7.1%.

Active management of diabetes requires measuring fasting and post-meal blood glucose levels. Average pre-diet values (143 mg/dL fasting and 175 mg/dL post-meal) reflected poor glucose tolerance (Table 3). At the time of the survey, the average blood glucose values were 99 mg/dL for fasting and 107 mg/dL for post-meal, which is a significant improvement.

It is clear in the literature that the most powerful predictor in a basic lipid panel of insulin resistance and cardiovascular health is the ratio of TG and HDL-cholesterol.^87,88,89,90 As low-density lipoprotein (LDL) cholesterol (which is calculated as the measurement of cholesterol mass within LDL particles) is made up of small (atherogenic) and large (healthy) particles, there is controversy about total LDL role as a predictor of cardiovascular events.⁷⁴ A low TG/HDL ratio reflects ‘pattern A’ (large LDL particles) and a high TG/HDL ratio reflects ‘pattern B’ (small LDL particles), the latter being more closely associated with both carbohydrate (particularly refined) intake and cardiovascular disease risk.^90,91 Because of its association with a less atherogenic lipid profile, a TG/HDL ratio close to 1.0 is regarded as predictive of low cardiovascular risk.⁹² A higher TG/HDL ratio is also associated with hyperinsulinaemia,⁹³ which is an independent risk factor for cardiovascular disease.⁹⁴ One concern about a low-carbohydrate diet is the typical slight rise in LDL cholesterol, although no study has shown its adverse outcome. Dietary interventions that lower LDL have also failed to show reductions in disease risk.^95,96 In fact, evidence has shown overall cardiovascular risk to improve on a well-formulated low-carbohydrate diet.⁹⁷ Higher LDL-C has been associated with longevity in one review of 68 000 patients.⁹⁸ Among study respondents who knew their values, after starting a low-carbohydrate diet, average HDL increased and average TG decreased. Most significantly, the pre-diet TG/HDL ratio was 3.5. After diet initiation, the average TG/HDL ratio was 1.4, which reflects a much lower cardiovascular risk.

Our survey showed marked reductions in medication use and costs, which is an important finding not only for consumers but also for insurance companies and taxpayers. While more than half of the respondents said drug costs were not relevant for them, a quarter of respondents were able to reduce their individual costs, and in those with a specific amount cited, the average reduction was $288 a month.

Decreased strength and difficulty in mobility are associated with poor health outcomes and increased mortality.^99,100,101 The ability to rise from the floor and have a strong walking speed is associated with healthy ageing.¹⁰² The number of survey respondents reporting moderate or severe mobility problems or being unable to walk declined from 278 to 49 individuals after diet initiation, and the number reporting moderate to severe pain declined dramatically after diet initiation, from 674 to 78. It is reasonable to conclude that the reported improvements in mobility were at least in part associated with reductions in pain. Chronic pain is a major reason why patients seek primary care, and our current approach to pain management has resulted in unintended consequences, contributing to the current opioid crisis.¹⁰³ Safe and natural methods for pain management need to be explored, and we are just beginning to learn the potential of nutritional therapy. This finding suggests further work needs to be conducted on dietary interventions for pain.

Similarly, anxiety and depression are common primary care symptoms.¹⁰⁴ Pharmaceutical management of these conditions has shown marginal benefit (confounded by significant misreporting of trials in the literature),^105,106 especially in those with mild or moderate symptoms where dietary or lifestyle interventions are likely to have a greater impact.^107,108 Antidepressant medications also carry significant side effects, especially in elderly patients already under a burden of iatrogenic polypharmacy.¹⁰⁹ Our survey respondents reported marked improvements in symptoms of anxiety and depression after starting a low-carbohydrate diet. These findings support low-carbohydrate nutrition as an alternative therapeutic approach for some mental health conditions and warrant continued research.¹¹⁰ How one feels between meals in mood and cravings is also critical to sustain any nutritional lifestyle plan. One well-known claim of low-carbohydrate diets is that one can lose weight and not be hungry.^111,112,113 Our survey respondents, more than one-third of whom had been on a low-carbohydrate diet over 2 years, strongly affirmed this. Table 10 shows the dramatic differences before and after low-carbohydrate diet initiation in self-reported symptoms of hunger, fatigue, poor concentration, mood swings, blood sugar swings, irritability and anxiety. Furthermore, participants also reported an increase in energy levels despite the lower caloric intake that naturally occurs with a low-carbohydrate diet.¹¹⁴

In this sample of successful, mostly long-term low-carbohydrate diet followers, the diet led to improvements in physical and psychological well-being for most of our study respondents. Feelings of emotional well-being are hypothesised to be important components for sustainable lifestyle change.¹¹⁵ These overwhelmingly positive changes suggest that a low-carbohydrate diet may be especially sustainable. In addition, 18% of participants reported no regular physical activity, supporting the idea that people do not need to exercise excessively to lose fat and improve metabolism.

Moreover, participants reported improvements in a variety of health conditions after initiating a low-carbohydrate diet, some of which have been studied before, at least preliminarily (improvements in symptoms related to migraines,¹¹⁶ irritable bowel syndrome,¹¹⁷ heartburn,¹¹⁸ polycystic ovary syndrome,¹¹⁹ NAFLD¹²⁰ and pain¹²¹), and some of which have not been studied well or much at all (e.g., libido, immune function and allergies). More research is needed on these topics.

This study was not without limitations. This was a retrospective study assessing a host of biomedical indicators reliant on self-reported data subject to recall bias. Although we believe that the questions we used are face valid, we did not use validated self-report scales. Therefore, some answers may not be as reliable and accurate as they would have been if we had used validated instruments. A detailed survey like this also lends itself to those who are highly engaged in the topic. Like any survey, the respondents may not be representative of all patients. Furthermore, in our appeal for participants, we did not differentiate between people who are followers of one of the many types of carbohydrate-restricted approaches such as low-carbohydrate, high-fat and low-carbohydrate, high-protein (although most of the practitioners who supported the survey favoured a low-carbohydrate, high-fat approach). As to the small sample of minorities, we feel this reflects, at least in part, the bias inherent in snowballing sampling. Many of the initial contacts when the survey was launched were not minority (non-white people comprise about 6% of the population of the state of West Virginia). As the survey was largely shared through social networks, it could be that race played a role in who was invited to take part.

Furthermore, this study lacked the rigour of a randomised, controlled trial. However, our goal was not to compare the results of a low-carbohydrate diet with some other dietary approach. Instead, we were interested in what people who were voluntarily choosing to follow some kind of carbohydrate-reduced nutritional approach would report about their lived experience.

The results of our survey support pursuit of other pertinent questions around myriad aspects of low-carbohydrate diets: long-term health effects, factors that facilitate success and overcoming of barriers, age-related differences in response to the diet and parameters of the optimal low-carbohydrate diet for overall health and well-being.