Low Carb Diets: Potential Short and Long-term Health Implications

Asia Pacific Journal of Clinical Nutrition 12(2003):397.
by Shane A Bilsborough MSc (Nutrition) and Timothy C Crowe PhD

Timothy C Crowe, PhD, is a Professor of Nutrition at Deakin University’s School of Health Sciences in Melbourne.

Low-carbohydrate diets for weight loss are receiving a lot of attention of late. Reasons for this interest include aplethora of low-carbohydrate diet books, the over-sensationalism of these diets in the media and by celebrities, andthe promotion of these diets in fitness centres and health clubs. The re-emergence of low-carbohydrate diets intothe spotlight has lead many people in the general public to question whether carbohydrates are inherently ‘bad’and should be limited in the diet. Although low-carbohydrate diets were popular in the 1970s they have resurgedagain yet little scientific fact into the true nature of how these diets work or, more importantly, any potential forserious long-term health risks in adopting this dieting practice appear to have reached the mainstream literature.Evidence abounds that low-carbohydrate diets present no significant advantage over more traditional energy-restricted,nutritionally balanced diets both in terms of weight loss and weight maintenance. Studies examiningthe efficacy of using low-carbohydrate diets for long-term weight loss are few in number, while few positivebenefits exist to promote the adoption of carbohydrate restriction as a realistic, and more importantly, safe meansof dieting. While short-term carbohydrate restriction over a period of a week can result in a significant loss ofweight (albeit mostly from water and glycogen stores), of serious concern is what potential exists for the followingof this type of eating plan for longer periods of months to years. Complications such as heart arrhythmias, cardiaccontractile function impairment, sudden death, osteoporosis, kidney damage, increased cancer risk, impairment ofphysical activity and lipid abnormalities can all be linked to long-term restriction of carbohydrates in the diet. Theneed to further explore and communicate the untoward side-effects of low-carbohydrate diets should be animportant public health message from nutrition professionals.

While the concept of low-carbohydrate diets have beenwith us for many decades, they appear to have had aresurgence in recent times and are currently generating awide-degree of public interest and media attention, fuelledby the rising tide of obesity and insulin resistance in thegeneral population. There are many variations on justwhat a ‘low-carbohydrate’ diet is. Some popular dietbooks such as Dr. Atkins’ Diet Revolution,1 TheCarbohydrate Addict’s Diet,2 Protein Power 3 and SugarBusters, 4 have common recommendations in that theyadvise consuming protein as the primary macronutrient forthe body, with the remainder of the energy to be made upfrom fat. Ad hoc restriction of carbohydrate may also bedone by using such simple rules as ‘no carbohydrates aftermidday’, as a way of self-restricting dietary carbohydrate.While some may classify the aforementioned diets as‘high-protein’ they are invariably hypoenergetic dietswhere energy from carbohydrate is restricted such that thepercentage contribution of energy from protein and fat israised.

Table 1 broadly characterises four generic diets used forweight loss in terms of macronutrient contribution. Thecommon thread amongst the diets listed in Table 1 is asimilar energy restriction, the primary reason for weightloss. While a ‘high-fat, low-carbohydrate’ diet doesseverely restrict carbohydrate; total protein and fat intakeare only raised by a small degree at the cost of a decreasein the overall nutritional adequacy of the diet.5 For thepurpose of this review a ‘low-carbohydrate’ diet is consi-deredto be a diet containing less than 100 g/d of carbo-hydratewith typical percentages by nutrient contribution toenergy being 50-60% for fat, less then 30% for carbo-hydrate,and 20-30% for protein. Furthermore, in order toadequately compare low-carbohydrate diets to conven-tionaldiets higher in carbohydrate a reference diet isassumed where carbohydrates contribute approximately

50% to the energy of the diet derived from an extensivevariety of unrefined wholegrain cereals, legumes, freshfruit, and vegetables (~200 g/day), as part of an overallhypo-energetic diet. For the aforementioned reference’high-carbohydrate diet’, protein is considered to representapproximately 20% of the energy with fat contributing thebalance.

Types of low-carbohydrate diets
The most widely used low-carbohydrate diet is the oneadvocated by Robert C. Atkins, M.D. His 1972 bookDr Atkins’ Diet Revolution 1 sold millions of copies withinthe first two years. His 1992 update, Dr Atkins’ New DietRevolution has sold over six million copies world wide.6Typically the diet involves four steps: a two-week‘induction’ period, during which the goal is to reducecarbohydrate intake to under 20 g/d to take the individualinto ketosis. During the induction phase, protein intakefrom such foods as beef, turkey, fish, chicken, and eggsare encouraged, although high protein intakes well abovethe individual’s habitual diet are not recommended -however,an unlimited consumption of fat is allowed.This phase of the diet allows no fruit, bread, grains,starchy vegetables or dairy products other than cheese,cream or butter. Carbohydrate restriction is lessenedduring the following stages until individuals determinethe level of carbohydrate they can consume whilemaintaining weight loss. For some, this level ofcarbohydrate restriction could be as low as 25 g/d and forothers it could be as high as 90 g/d. The diet alsoencourages the person to check their urine for ketonebodies to ensure that ketosis is maintained.

The Carbohydrate Addict’s Diet 2 aims to breakcravings for ‘fat-causing carbohydrates’ by limiting theintake of carbohydrate-containing foods to only one mealper day. The Protein Power diet 3 provides for 0.75 g/d ofprotein per kilogram of body weight with less than 30 g ofcarbohydrates per day allowed in the induction phase andup to 55 g/d thereafter. The Protein Power diet is, by itsnature, a carbohydrate- and protein-restrictive diet whencompared against habitual intakes of carbohydrate andprotein of most individuals. The Sugar Busters! diet 4advises the avoidance of sucrose and high-glycaemicindex foods such as potatoes, pasta, corn, white rice, andcarrots. According to the Sugar Busters! diet, it is high-glycaemicindex foods that cause a spike in insulinresponse that are responsible for the deposition of fat andthe cause of insulin resistance (the fact that sucrose has amedium glycaemic index seems to be ignored in therationale behind the diet). The main premise of mostpopular low-carbohydrate diet books is that a reduction incarbohydrates will result in a lowered basal insulin leveland hence promote more triglyceride lipolysis into freefatty acids - leading to fat loss in overweight people. Inreality, these diets work in the short term as they result ina reduction in total kilojoule intake which inherently leadsto weight loss due to an energy deficit.

Weight loss achieved on low-carbohydrate diets
In one of the few published papers on the Atkins’ diet, 18subjects who followed the dietary regimen lost 7.7 kg in 8weeks.7 The dietary protocol represented an averagekilojoule restriction of approximately 4500 kJ/d relativeto the subjects’ average previous isoenergetic diet. Thedieters lost on average 3.6 kg in the first two-weekinduction period. In the following three two-weekperiods, weight loss was in the order of 1.4 kg perfortnight. The results showed that weight loss wasentirely due to energy restriction achieved through adecrease in carbohydrate consumption by 90% while theactual amounts of fat and protein eaten changed littlecompared to the habitual intake of the participants prior tothe commencement of the diet.

Comparing ketogenic and non-ketogenic hypo-energeticdiets for weight loss in a one-month studyshowed no statistical advantage of either diet over theother in terms of weight loss.8 These aforementionedstudies confirmed many previous studies, whichconcluded that lower energy intakes rather than lowcarbohydrate intakes may be the key to successful weightloss.9 The US National Weight Control Registry, whichcompiles details of individuals who have lost more than

13 kg for a year or more, analysed the diets of 2,681members.10 They found that fewer than 1% of thesesuccessful people had followed a diet classified as ‘lowcarbohydrate’ (defined as 24% or less total dailykilojoules coming from carbohydrate) suggesting that thistype of diet is not realistic for the achievement of long-termweight loss. Further to this, a study analysing thediets of over 10,000 free-living adults in the UnitedStates 11 found that those individuals consuming a dietwith 55% of energy from carbohydrates, compared tothose whose diets comprised zero to 30% energy contri-butionfrom carbohydrates, actually ate the same amountof food in terms of weight (2,271 ± 21 g versus 2,225 ±

103 g), however consumed less energy (7,728 ± 71 kJversus 8,530 ± 307 kJ), more fibre (18 ± 3 g versus 9 ±

0.4 g), and less fat (52 ± 1 g versus 101 ±3g). Individualsconsuming a greater percentage of carbohydrates ate morelow-fat foods, unrefined grain products and fruits andthese adults were more likely to have a BMI less than 25kg/m 2 . Such data suggests that a high-carbohydrate diet,which contains unrefined grains, cereals, and plenty offibrous matter is more nutritionally adequate than low-carbohydratediets and of a lower energy density whichmay be important in weight control.

One advantage that has been attributed to low-carbohydratediets over conventional diets is the potentialhigher satiety value of protein and fat and the anorecticeffect of ketosis in helping to enhance a decrease inappetite. Many factors however are able to influenceappetite, hunger and subsequent food intake, withmacronutrient intake being just one of these factors, hencethe issue of satiety in dietary restriction is an area of greatcontroversy. In terms of the greatest degree of satiety forcontrolling nutrient intake, fat-restricted diets aregenerally regarded as optimum with the efficacy ofweight loss on such diets attributed to low energydensity.12,13 Further studies also suggest that diets wherethe macronutrient contribution is modified wherecarbohydrate contributes approximately 40% to energyand fat to 36% of energy (fat source mostly mono-unsaturated)can also achieve realistic weight loss andallow a greater chance of long-term compliance.14,15 Thechance of long-term compliance to any diet will mostlikely result from short-term weight-loss strategies thathave the best chance to become habitual rather thanrestriction of macronutrient consumption per se.

Potential short-term health implications

One of the common metabolic changes seen when aperson follows a low-carbohydrate diet is ketosis. Whendietary carbohydrates are in limited supply, the body willutilise its reserves of glycogen in order to meet glucosedemands. Glycogen stores in the body are quite smallwith approximately 70-100 g in the liver and 400 g inmuscle. Most of these glycogen stores are exhaustedwithin 24 to 48 h of carbohydrate restriction. As eachgram of glycogen is bound with 3 g of water, then asimple calculation shows that a ‘weight loss’ of around 1-2kg can be achieved within the first week of the diet,albeit due to diuresis and not to the burning of fat stores.Loss of glycogen and water is not a true measure ofweight loss as glycogen and water stores will bereplenished once the diet ends. The diuretic effect of alow-carbohydrate intake is limited to the first week of thediet, with ongoing weight loss after this time entirely dueto the laws of energy balance.16

Once glycogen stores are depleted, the body begins toincrease fat oxidation as a means of meeting the majorityof its energy demands that can not otherwise be met bygluconeogenesis (the production of glucose from certainamino acids) and glycerol liberation from triglyceridebreakdown. Fatty acids liberated into the blood are ableto be oxidised by the liver for energy production. Fattyacids can also be partially oxidised by the liver to formacetoacetate, which can then be further converted to â -hydroxybutyricacid, both of which are collectivelytermed ketone bodies. Ketone bodies are able to be usedby all tissues containing mitochondria, which includesmuscle and brain. Ketone bodies themselves are filteredby the kidneys and cause an increased renal loss ofsodium, thereby increasing water loss.16 Currently, thereis no consensus as to what is the absolute cut-off limit forthe maximal amount of carbohydrate intake necessary toinduce ketosis. Induction of ketosis is likely to vary on anindividual basis, however, intakes in the range of 50-100g of carbohydrate a day are generally reported - typicallyrepresenting less than 20% of energy from carbohydrate.17

Studies examining the long-term safety of ketogenicdiets are few in number with most of the available datacoming from the application of ketogenic diets in thetreatment of paediatric epilepsy.18 The diet used in thispatient group is a high-fat, adequate protein, low-carbohydratediet designed to mimic the biochemicalchanges that occur during starvation. Studies of childrenwho have followed a ketogenic diet for management ofepilepsy found that about 50% of children will continueon the diet for at least a year.18 Reasons for discontinuingthe ketogenic diet were due to either a lack of efficacy ordue to the restrictive nature of food choices. Commonadverse events attributed to the diet included dehydration,gastrointestinal symptoms, hypoglycaemia, as well ascarnitine and vitamin deficiencies. Cognitive effects,hyperlipidaemia, impaired neutrophil function, uroli-thiasis,optic neuropathy, and osteoporosis have also beenreported to occur in some patients following ketogenicdiets.19 In addition, elevation of blood uric acid levels isa well-recognised side effect of prolonged ketosis.7Long-term effects of exposure of the body to elevated uricacid levels while on ketogenic diets have not beenstudied, but potential does exist for arthritic and renalcomplications due to long-term hyperuricaemia. Grossstudies on a normal adult population using a ketogenicdiet for weight loss for periods of a year or more (asdescribed in the aforementioned paediatric group) havenot been performed though the documented side-effectsmay potentially pose a serious health risk to individuals.

Dietary adequacy
All hypoenergetic diets result in loss of body weight andbody fat. Losses of protein and fat are the same during aketogenic diet as during a hypoenergetic, non-ketogenicdiet hence no diet is superior to another in terms ofpreservation of lean body mass.20 However, low-carbohydratediets are at greater risk of being nutri-tionallyinadequate as they enforce restriction of foodchoices. Typically, low-carbohydrate diets are low infibre, thiamin, folate, vitamins A, E, and B6, calcium,magnesium, iron, and potassium.21 In the absence ofsupplemental multivitamins, there is a real risk of nutri-tionaldeficiencies occurring. Low-carbohydrate diets arealso usually higher in saturated fat and cholesterol withprotein mainly being derived from animal sources.21Comparison of a range of popular diets revealed that low-carbohydratediets (defined as less then 30% of energyfrom carbohydrate) fared worse in terms of dietaryadequacy (as defined by the amount of inclusion of thefive major foods groups and alignment with the U.S.Dietary Guidelines) while high-carbohydrate diets(defined as greater than 55% of energy fromcarbohydrates) gave the highest dietary adequacy score.5The aforementioned study gave a rating, known as theHealthy Eating Index (HEI) to the discussed dietarytypes: high-carbohydrate diets received a HEI score of

82.9 (highest possible score is 100) which was the highestscore recorded for the range of diets analysed while low-carbohydratediets received a score of 44.6 (the lowestrecorded score in the study).

Physical activity
Short-term effects of low-carbohydrate diets oftenreported by individuals include nausea, thirst, polyuria,headache, dizziness, halitosis and fatigue.21 Dehydrationis very common during periods of restricted carbohydrateconsumption due to increased water loss associated withketotic-induced diuresis and water loss from depletion ofglycogen stores. Substantial losses of water through sweatduring exercise can lead to further losses of water andelectrolytes.22 Commencing any form of physical activitywhilst in a dehydrated state, whether it be an elite athleteor the average person on a weight loss diet, is known toimpair performance and metabolic function and havenegative effects such as; an early onset of fatigue,impairment of mental function, increase in bodytemperature, changes in blood pressure, and heatstroke.22,23 Hydration prior to commencement of, andduring physical activity, should be an area of high priorityfor all individuals, but especially those that may befollowing a carbohydrate-restricted diet.

As an increase in physical activity is often undertakenin conjunction with dieting, the influence of carbohydraterestriction on physical performance is an important issueto examine. High-intensity anaerobic exercise requirescreatine phosphate breakdown as well as muscle glycogenutilisation.24 It has been shown that low-carbohydratediets reduce mean power output due to a reduction inmuscle glycogen stores and hence a decreased rate ofglycolysis.25 Several studies have shown that the timeuntil the onset of fatigue during high-intensity exercise(defined as at, or close to, 100% VO2max) is accentuatedin untrained subjects who had been following a low-carbohydratediet prior to exercise when compared toindividuals following moderate- or high-carbohydratediets.26-30 However, similar studies to those just describedfound that in subjects defined as highly trained orphysically active, onset of fatigue when following a low-carbohydratediet was not significantly differentcompared to conventional diets.27-29,31 A lowcarbohydrate intake over a period of three to four daysresults in metabolic acidosis due to increased circulationof free fatty acids and 3-hydroxybutyrate concentrations.It has been suggested that the presence of metabolicacidosis may be a factor in leading to fatigue after theadoption of the afore-mentioned dietary regime,32however few studies support this proposal.

In aerobic exercise, muscle glycogen is an importantsubstrate for ATP production within contracting skeletalmuscle. The importance of ATP availability is demon-stratedby the fact that fatigue is often associated withmuscle glycogen depletion.33 The significance of carbo-hydrateavailability during moderate-intensity exercise(60-70% VO2max) is demonstrated by the earlier onset offatigue once glycogen stores are depleted as the body isunable to oxidise fat at sufficient rates to meet energyrequirements. Work rates are hence reduced to lowerlevels of intensity (30-50% VO2max) to compensate forthe reduced rate of energy production.34 While littlepublished data exists examining the effects of carbo-hydraterestriction on the physical performance of averageuntrained individuals, there is evidence to suggest thatsuch individuals do have a diminished fat oxidative capa-cityhence have a greater reliance on glucose as fuelsource during exercise.35 Taking these findings intoconsideration, it is worth to pose the question as to whateffect carbohydrate restriction may have on the physicalperformance of inactive, overweight and untrainedindividuals and if such dietary regimens may compromiseweight loss due to impairment of the ability to exercise atthe individuals maximal potential. Such studies at presenthave not been performed.

Potential long-term health implications

Insulin response
One of the main rationales promoted by advocates of low-carbohydratediets is that insulin secretion, in response tocarbohydrate ingestion, is the cause of the metabolicimbalance that promotes obesity. Carbohydrate ingestionresults in increased insulin secretion with subsequentglucose uptake by cells and inhibition of lipolysis. If thepost-prandial secretion of insulin could somehow beattenuated by restriction of carbohydrate-containing foodsthen this would promote a more favourable environmentfor fat breakdown and hence, as advocated by supportersof low-carbohydrate diets, weight loss. While such asituation is true in theory it is in fact overly simplistic andthe actual reality is much more complex. What is knownis that energy restriction, irrespective of dietary compo-sition,promotes weight loss and improvement of gly-caemiccontrol.21 A study by Golay et al.,20 in humansubjects demonstrated that an intake of either 15% or 45%carbohydrate as part of a total daily energy intake of 4.2MJ over 6 weeks produced no significant difference inweight loss, (8.9 ± 0.6 and 7.5 ± 0.5 kg respectively)despite significantly different insulin levels of 57.6 ± 0.6and 88.2 ± 9.6 pmol/L respectively. To focus on oneparticular metabolic response, that being post-prandialcarbohydrate-stimulated insulin secretion, presents a veryunbalanced view of the complex metabolic changesleading to obesity.

What is overlooked in the simple metabolic situationpromoted by proponents of low-carbohydrate diets is thatdietary amino acids are also able to stimulate insulinsecretion without augmenting glucose concentrations.36-39Investigations examining the insulin response to wholefoods showed that protein foods such as meat and fishelicited a greater peak insulin concentration than whitepasta.36 The effect that protein can have on insulinsecretion is an interesting finding as it raises several areasof concern, primarily what the rise in insulin secretionmay have in terms of glycaemic control, particularly inindividuals with diabetes or impaired glucose tolerance.In one long-term study of the metabolic effects of high-proteindiets, subjects were fed either 1.87 g/kg/d (highprotein) or 0.74 g/kg/d (normal protein) for 6 months.37It was found that the glucose-stimulated insulin responsein the high-protein group was significantly increasedwhen compared to the normal protein group. Intere-stingly,insulin levels in the high protein group remainedelevated for up to eight hours in the postprandial state. Ina recent review of the glycaemic response of foods as apredictor of insulin response it was concluded that theglycaemic response accounted for only 23% of thevariability in insulinaemia.40

Chronic ingestion of a low-carbohydrate diet, coupledwith elevated amounts of protein as a major macro-nutrientresults in increased hepatic glucose productionand decreased peripheral glucose utilisation, both indi-cativeof an insulin resistant state.41,42 Although morestudy in this area is needed, early findings suggest thatincreased basal insulin release, higher fasting-glucoseproduction and enhanced gluconeogenesis all increase thedemand on insulin release from the pancreas and mayhasten the onset of diabetes in susceptible individuals dueto pancreatic beta cell failure.37,38 In contrast, the use ofhigh-protein, hypo-energetic weight-loss diets (asopposed to high-protein, isoenergetic diets) in themanagement of type 2 diabetes has shown comparableresults to reduced-fat diets in terms of improvements inmetabolic para-meters such as blood glucose, lipids,insulin response as well as overall weight loss.39

Cardiovascular complications
Any diet that results in weight loss will elicit a favourableresponse on blood lipid parameters.44,45 Typical reportedchanges include a reduction in total cholesterol, trigly-cerides,LDL- and HDL-cholesterol while free fatty acidsare elevated. However, nutrient composition of energy-restricteddiets can have differing effects on absolutechanges in blood lipids. Greater decreases in LDL-cholesterolare seen during active weight loss when dietsare low in saturated fat, however a low-carbohydrate dietis typically higher in saturated fat than conventionalweight-loss diets. A review of the effect of the use ofisoenergetic ketogenic diets on blood lipids found that,overall, LDL-cholesterol and total cholesterol tended tobe elevated while HDL-cholesterol levels were lowered.21As LDL-cholesterol is considered a major contributor tothe process of atherogenesis, then long-term use of low-carbohydratediets, be it for weight reduction or weightmaintenance, may have the potential to put an individualat greater risk of heart disease. Proponents of low-carbohydratediets often point to the fact that high-carbohydratediets increase plasma triglycerides.46 Theclinical significance of carbohydrate-induced hypertri-glyceridaemiain individuals who are otherwise healthy isan area of great debate and currently no firm conclusionscan be made. In many persons however, the effect ofcarbohydrate feeding on triglyceride synthesis can besignificantly diminished by weight loss, exercise, anddietary restriction irrespective of the macronutrient make-upof the diet.47

While important tissues such as the brain, muscle, andheart can utilise ketone bodies as a primary fuel sourceduring carbohydrate restriction, there may be a metaboliccost associated with the use of fatty acids as a fuel due tothe inability of the body to replenish key intermediates ofthe citric acid cycle necessary for energy production. Onestudy performed in isolated rat hearts found that when theheart was perfused in a solution containing primarilyacetoacetate (a ketone body) as a fuel source, contractileability of the heart was reduced by 50% within 60minutes.48 This contractile failure was reversed by theaddition of pyruvate, which can be synthesised fromglucose. Such results demonstrate that ketone bodies arenot a self-sufficient fuel for the working heart, but requireaugmentation from a carbohydrate-derived substrate sucha pyruvate.

A study by Best et al.,49 investigating potential cardiaccomplications in 20 paediatric patients on a ketogenic dietfor management of epilepsy found cardiac abnormalities(as defined by changes in the QT interval of an ECG) inthree of the patients on the diet. Interestingly, there was asignificant correlation between prolonged QT interval andlow serum bicarbonate (suggesting increased bloodacidity) and high beta-hydroxybutyrate (levels of whichare elevated in ketosis). Prolongation of the QT intervalis of clinical importance due to an increased risk forventricular dysrhythmia and sudden death.49,50 Suchchanges in the QT interval have also been observed inanorexia nervosa and very-low-kilojoule diets forobesity;51 both metabolic states where ketosis is verylikely to occur. Ketogenic diets result in an increase inplasma free fatty acids due to increased lipolysis whileprolonged elevation of free fatty acids has been linkedwith cardiac arrhythmias.52 Over 60 deaths in individualson medically supervised very-low-kilojoule liquid proteindiets have been reported 53 with the suggestion that cardiacarrhythmias were the cause of death.54 Further studieshowever have suggested that low-energy diets withadequate supervision are safe and produce no ECGabnormalities,55 while other studies suggest patients whoingest a nutritionally balanced low-kilojoule diet overseven days were also void of cardiac abnormalities.56Considering that both starvation diets 57,58 and starvation-likestates such as anorexia nervosa can induce fatalcardiac dysrhythmias and both states can induce ketosisthen the safety of low- and very-low-carbohydrateketogenic diets needs to be explored further, especially inlight of the lack of information on electrolyte andphysiologic changes during such dieting regimes.

Bone health
A potential effect of low-carbohydrate diets on bonehealth is an important consideration. The loss of thebody’s calcium stores is a major concern, especially inwomen and the elderly, as it is strongly linked toosteoporosis.59,60 Observational studies and controlledtrials with children, young adults, and the elderly allsupport the important role for calcium intake in buildingand maintaining bone mass and reducing bone loss.59,60Low-carbohydrate diets promote the restriction of dairyproducts, particularly milk and yoghurt, which are themain sources of calcium in the diet. As peak bone mass isan important factor in determining long-term fracture andosteoporosis risk, adoption of dieting practices thatrestrict calcium intake (particularly in those under the ageof 30) have the potential to compromise the attainment ofpeak bone mass. Individuals with the highest peak bonemass after adolescence have the greatest protective advan-tagewhen the inexorable declines in bone density asso-ciatedwith increasing age, illness, and diminished sex-steroidproduction take their toll.61 As low-carbohydratediets are known to be deficient in calcium 21 then theadoption of this type of dieting practice, especially iffollowed as a long-term eating strategy in those under 30years of age, pose a real possibility of increasing osteo-porosisrisk later in life.

Low-carbohydrate diets have the potential to generatea sub-clinical chronic metabolic acidosis (via the presenceof ketone bodies in blood) which can then promotecalcium mobilisation from bone.62-65 Blood acidificationcan increase glomerular filtration rate and decrease renaltubular reabsorption of calcium with a concomitantincrease in activity of osteoclasts and inhibition ofosteoblasts, further increasing bone resorption. Studieshave shown that by increasing the total proportion ofprotein consumed as meat, which contains sulphur aminoacids (namely methionine and cysteine), in the dietincreases calcium, potassium, sodium and ammonia lossin the urine due to a change in blood acidity.66 Low-carbohydratediets promote the consumption of animalprotein as a good way of limiting carbohydrate intakehence compounding the potential effects on blood acidityand calcium loss. Barzel and Massey 62 proposed thatdiets with the potential to increase renal acid load lead tocalciuria which can have adverse affects on bone unlessbuffered against the consumption of alkali-rich foods suchas fruits and vegetables. Recent studies though appear tocontradict this theory with a large-scale epidemiologicalstudy of over 900 adults showing no negative effect ofconsumption of animal protein on bone mineral density.67Conflicting reports on the effect of animal protein andbone health could be due in part to differences in parti-cipantages, calculalation of protein intakes, dietary datacollection problems, and anatomic sites evaluated. Dietsbased mainly on plant proteins do not appear to augmentcalcium loss, an effect most likely attributed to a higherphosphate intake (dietary phosphate has the ability toincrease renal tubular reabsorption of calcium) and alower intake of sulfur-containing amino acids.68 Studieshave not been performed on the adverse aspects of low-carbohydratediets on bone health in adults, but thepotential metabolic mechanism to induce, or at leasthasten, osteoporosis do exist.

Cancer risk
There is overwhelming evidence for a protective effect offruits and vegetables in almost all major cancers afflictingwestern society today including colorectal, breast,pancreatic, lung, stomach, oesophageal, and bladdercancer.69,70 Fruits and vegetables contain a vast array ofcompounds that are implicated in providing protectionagainst cancer. For examples, such substances such asantioxidants, fibre, isothiocyanates (in cruciferous vege-tables),allyl sulphides (in onions and garlic), flavonoids,and phenols have all been linked to augmenting thebody’s protective mechanism against cancer promotion.The nature of a low-carbohydrate diet however is one thatis low in fruits, vegetables (if starchy vegetables aren’tadequately replaced with other types of low-carbohydrate-containing vegetables) and grains thuspotentially placing an individual at an increased cancerrisk if the diet is followed long term. Furthermore, theevidence strongly suggests that it is not the consumptionof one or two varieties of vegetables and fruit that conferbenefit, but rather the intake of a wide-variety of plantfoods (the latter being a common factor in those who havea lower risk of cancer).

The potential link between increased intakes of meat(typically seen on low-carbohydrate diets) and bowelcancer risk can not be ignored as it has been suggestedthat the link between meat and the consumption of animalprotein with cancer is as strong as the association of fatwith cancer.71 In a recent review of prospective cohortstudies on meat consumption and colorectal cancer risk, itwas found that daily increases of 100 g of all meat or redmeat is associated with a significant 12-17% increasedrisk of colorectal cancer.72 Epidemiologic evidence oncolorectal cancer risk and meat consumption from 32case-control and 13 cohort studies supported the hypo-thesisthat meat consumption (in particular red meat), isassociated with a modest risk of increased colorectalcancer risk.73 Not all the evidence however is equivocal 74and it has been suggested that the epidemiological dataare much more consistent with there being a protectiverole of fruit, vegetables and whole-grain cereals and norole for meat in colorectal cancer.75 A low-carbohydratediet promotes increased meat consumption at the expenseof fruit and vegetable intake.

Many mechanisms have been proposed for theassociation between red meat and colon cancer such aslow intakes of protective dietary factors such as fibre anda displacement of protective compounds found in fruitand vegetables in the diet.72,73 At present, however, nolong-term consumption data exists to assess cancer risk inindividuals following low-carbohydrate diets. A signi-ficantreduction in dietary fibre is typically seen when anindividual follows a low-carbohydrate diet and this mayexplain reports of constipation in people following thesetype of diets 21 as dietary fibre is only found in foods ofplant origin such as cereals, fruits and legumes. Dietaryfibre can have a myriad of benefits in the colon such asdiluting carcinogenic compounds, increasing stool transittime, production of beneficial fermentation products suchas butyric acid and a lowering of pH, all of which havebeen proposed as being protective against colon cancer.76

Practical advice to health professionals
In terms of weight loss, the simple fact of energy balancecan not be ignored; any diet that is hypoenergetic willresult in weight loss. When using loss of body fat as atrue measure of weight loss, then low-carbohydrate dietspresent no significant advantages to the dieter over nutri-tionallybalanced, hypoenergetic diets. Based on theavailable evidence, the reverse may in fact be true, in thatlong-term compliance to a low-carbohydrate may put anindividual at greater risk of an array of metabolic diseaseswithout the achievement of sustainable weight loss. Acomprehensive recent review of popular diets concludedthat a diet that is high in fruit and vegetables, whole-grains,legumes, and low-fat dairy products as well asbeing moderate in fat and kilojoules will result in thegreatest chance of weight loss and maintenance.21 Suchdiets are also associated with fullness and satiety and canreduce risk of chronic disease. Low-carbohydrate dietsachieve very few of these aims. Certainly within thescope of the ideal type of diet to follow for long-termweight loss there is considerable scope for variation in theactual advice given. Diets can be recommended to suitthe individual’s own food preferences within the guide-linesgiven previously. For some individuals this advicemay be to have a higher protein intake, combined with amoderate reduction in fat and carbohydrate, while forothers low-fat, high-carbohydrate food choices may beappropriate.

Advice should be given to an individual following alow-carbohydrate diet to help avoid some of the potentialmetabolic consequences known to be associated with thisdiet. For example, advice would include: to increase theintake of water to help prevent dehydration; ensure anadequate intake of fibre from non-starch containing foods;and to consume an adequate amount of calcium eitherfrom the consumption of low-fat dairy products, cannedfish with bones or from the use of supplements. The useof a general multivitamin formulation would also seemprudent in light of the array of vitamin and mineraldeficiencies that may potentially exist. Certainly thosewith a history of heart problems should be stronglydissuaded from restricting carbohydrates whilst under-takingvigorous exercise due to potential cardiac abnor-malitiesassociated with ketogenic diets.Although this review has focused on the contributionof free fatty acids and ketone bodies to service the body’senergy demands, the contribution of gluconeogenesisfrom the bodies amino acid stores has been estimated at

55% of endogenous glucose production in normalsubjects 5 to 12 hours after a large evening meal thuspotentially increasing muscle breakdown.77 Furtherstudies in this area would be useful to determine whatminimal amount of dietary carbohydrate is enough toreduce gluconeogenesis, and thus reduce the demand onskeletal and cardiac muscle protein stores. Currently it isestimated that ~120 to 180 grams of carbohydrate isnecessary for inactive individuals to consume to avoidcatabolism of bodies protein stores to fuel glucoseproduction.78 Weight gain at the cessation of a low-carbohydratediet needs also to be addressed. Althoughdata on this issue is scarce, mechanisms for weight gaincould be related to the up regulation of glycogen synthaseactivity causing a super-compensation of muscle and liverglycogen as in an athlete’s carbohydrate-loading phase,79also resulting in greater hydration than pre-diet hydrationstatus. This resultant weight gain is also important tounderstand from a health professional’s point of viewbecause the average person (who has been on a low-carbohydratediet) may misinterpret weight gain aftercarbohydrates are increased in the diet as carbohydratesturning to fat’ thus propagating the perception of thenegative role of carbohydrates in the diet.

The potential message that low-carbohydrate dietssend out into the public arena is ‘don’t eat carbohydratesbecause they’ll make you fat’. This concept is clearlyillustrated from a quote from Dr. Atkins “A carbohydrate-restricteddiet is so effective at dissolving adipose tissuethat it can create fat loss greater than occurs duringfasting” and “For many of us, the bypassing of carbo-hydratesis our ultimate solution”.1 Many health clubs,TV programs, gymnasiums and personal trainers advocatea ‘No carbohydrates after midday’ policy as theirinterpretation of a low-carbohydrate diet. The authors’own studies, based on personal involvement in the healthand fitness industry in Australia, suggest that well over

200 gymnasiums Australia-wide are advocating orpromoting some variation of a low- or no-carbohydratedietary regime (unpublished observations). It has beenclearly demonstrated, however, that the conversion ofcarbohydrate to fat via de novo lipogenesis is not a majorpathway in humans.80 It has also been repeatedly shownin studies that it is the overall kilojoule content of the dietthat dictates the magnitude of fat loss, not the nutrientcomposition.

With the promotion of the purported ‘positive weight-losseffects’ of low-carbohydrate diets in the media forthe past decade, it is important that the general public areaware of the potential negative effects of this type ofdieting practice. The marketing of celebrities such asGeri Halliwell and other Hollywood stars who have lostweight on these diets does little in the way to promotesafe and realistic dieting practices. At the beginning of

2001 the British Dietetic Association called for a stop tocelebrity-endorsed health and weight-loss claims that donot stand up to scientific scrutiny.81 Many people in thecommunity feel that carbohydrates are inherentlyunhealthy due to their perceived major role in promotingobesity. Nutrition health professionals need to correctthis misconception with balanced, factual and realisticadvice.

In the face of the rising tide of obesity in developedcountries, the lure of easily attainable weight loss byfollowing a low-carbohydrate diet is certainly appealing,yet little discussion is given to the potential negativehealth aspects potentially associated with this type ofdieting regime. When health professionals speak of low-carbohydratediets they often repeat the message of lackof concentration, kidney problems and bad breath andoften say ‘low-carbohydrate diets can be dangerous’.This does little to deter people in the quest for weight lossand there is the commonly held belief that ‘As long as Ilose weight I don’t care what I have to do.’ By deliveringa stronger message in the future and addressing suchserious potential health aspects to low-carbohydrate dietssuch as potential cardiac complications, osteoporosis,muscle loss and possibly insulin resistance, people arebetter able to make informed choices based on the latestscientific thinking about the risks associated with populardietary practices such as low-carbohydrate diets. Futurestudies are certainly warranted, especially in theexamining of the long-term efficacy of the use of low-carbohydratediets and how much this type of dietingpattern may alter an individuals perception of foods andfood choices well into the future.


  1. Atkins RC. Dr Atkins’ New Diet Revolution. New York,NY: Avon Books; 1992.

  2. Heller RF, Heller RF. The Carbohydrate Addict’s Diet.New York: Penguin books; 1991.

  3. Eades MR, Eades MD. Protein Power. New York, NY:Bantam Books; 1996.

  4. Steward HL, Bethea MC, Andrew SS, Balart LA. SugarBusters! New York, NY: Ballantine Books; 1995.

  5. Kennedy ET, Bowman SA, Spence JT, Freedman M, KingJ. Popular diets: Correlation to health, nutrition, andobesity. J Am Diet Assoc 2001:101:411-20.

  6. Atkins RC. Dr. Atkins’ New Diet Revolution. New York,NY: Avon Books; 1998 Revised and updated version.

  7. Larosa J, Fry A, Muesing R, Rosing D. Effects of high-protein,low-carbohydrate dieting on plasma lipoproteinsand body weight. J Am Diet Assoc 1980; 77: 264-270.

  8. Wing R, Vazquez J, Ryan C. Cognitive effects ofketogenic weight reducing diets. J Obes Relat MetabDisord 1995; 19: 811-6.

  9. Rabast U, Vornberger K, Ehl M. Loss of weight, sodiumand water in obese persons consuming a high or lowcarbohydrate diet. Ann Nutr Metab 1981; 25: 341-9.

  10. Wyatt HR, Seagle HM, Grunwald GK, Bell ML, KelmML, Wing RR, Hill Jo. Long-term weight loss and verylow carbohydrate diets in the National Weight ControlRegistry. Obesity Research 2000; 8 Suppl 1: 87S.

  11. Bowman S, Spence J. A comparison of low-carbohydratevs high-carbohydrate diets: Energy restriction, nutrientquality and correlation to body mass index. J Am Coll Nut 2002; 21: 268-74.

  12. Blundell JE, Lawton CL, Cotton JR, Macdiarmid JI.Control of human appetite: implications for the intake ofdietary fat. Ann Rev Nutr 1996; 16: 285-319.

  13. Rolls BJ. Carbohydrates, fats and satiety. Am J Clin Nutr 1995; 61: 960S-967S.

  14. Walker KZ and O’Dea K. Is a low fat diet the optimal wayto cut energy intake over the long-term in overweightpeople? Nutr Metab Cardiovac Dis 2001;11:244-8.

  15. Walker KZ, O’Dea K, Nicholson GC, Muir JG. Dietarycomposition, body weight, and NIDDM. Comparison ofhigh-fiber, high-carbohydrate, and modified-fat diets.Diabetes Care 1995; 18: 401-3.

  16. Denke M. Metabolic effects of high-protein, low-carbohydratediets. Am J Cardiol 2001; 88: 59-61.

  17. Shils ME, Olson JA, Shike M. eds. Modern nutrition inhealth and disease. Philadelphia: Lea & Febiger; 1994, p.996.

  18. Freeman JM, Vinning E, Pillas DJ, Pyzik P, Casey J,Kelly M. The efficacy of the ketogenic diet-1998: Aprospective evaluation of intervention in 150 children.Padiatrics 1998; 102: 1358-63.

  19. Tallian K, Nahata M, Tsao CT. Role of the ketogenic dietin children with intractable seizures. Ann Pharmacother 1998; 32: 349-61.

  20. Golay A, Allaz A-F, Morel Y, de Tonnac N, Tankova S,Reaven G. Similar weight loss with low- or high-carbohydratediets. Am J Clin Nutr 1996; 63: 174-8.

  21. Freedman MR, King J, Kennedy E. Popular diets: ascientific review. Obesity Research 2001;9 Suppl 1:1S-40S.

  22. Coyle E, Montain S. The influence of graded dehydrationon hyperthermia and cardiovascular drift during exercise.J Appl Phys 1992; 73: 1340-50.

  23. Hargreaves M. Fluid and energy replacement for physicalactivity. Aust J Nut Diet 1997;53 Suppl 4:1-48.

  24. McCully K, Vandenborne K, DeMeirleir K, Posner J,Leigh J. Muscle metabolism in track athletes using 31 Pmagnetic resonance spectroscopy. Can J PhysiolPharmacol 1992; 70: 1353-9.

  25. Haff G. Roundtable discussion: low-carbohydrate dietsand anaerobic athletes. Strength and Conditioning Journal 2001; 23: 42-61.

  26. Greenhaff P, Gleeson M, Maughan R. Diet inducedmetabolic acidosis and the performance of high intensityexercise in man. European Journal of Applied Physiology 1988; 57: 583-590.

  27. Greenhaff P, Gleeson M, Maughan R. The effect ofdietary manipulation on blood acid-base status and theperformance of high intensity exercise. European Journalof Applied Physiology 1987;56:331-337.

  28. Greenhaff P, Gleeson M, Whiting P, Maughan R. Dietarycomposition and acid base status: Limiting factors in theperformance of maximal exercise in man? EuropeanJournal of Applied Physiology 1987; 56: 444-450.

  29. Maughan R, Poole D. The effect of a glycogen-loadingregime on the capacity to perform anaerobic exercise.European Journal of Applied Physiology 1981; 46: 211-219.

  30. Balsom P, Gaitanos G, Soderlund K, Ekblom B. High-intensityexercise and muscle glycogen availability inhumans. Acta Physiol Scand 1999;165: 337-345.

  31. Hargreaves M, FinnJ, Withers R, Halbert J, Scroop G,Mackay M, Snow R, Carey M. Effect of muscle glycogenavailability on maximal exercise performance. Eur J ApplPhysiol Occup Physiol 1997;75 (2):188-192.

  32. Maughan R, Greenhaff P, Leiper J, Ball D, Lambert C,Gleeson M. Diet composition and the performance ofhigh-intensity exercise. J Sport Sci 1997;15:265-275.

  33. Coggan A, Coyle E. Reversal of fatigue during prolongedexercise by carbohydrate infusion or ingestion. J ApplPhysiol 1990; 68: 990-6.

  34. Coyle E. Substrate utilization during exercise in activepeople. Am J Clin Nut 1995; 61: 968S-979S.

  35. Kiens B, essen-Gustavsson B, Christensen N, Saltin B.Skeletal muscle substrate utilisation during submaximalexercise in man: effect of endurance training. J Physiol 1993; 469: 459-478.

  36. Holt S, Brand-Miller J, Petocz P. An insulin index offoods: the insulin demand generated by 1000-kJ portionsof common foods. Am J Clin Nutr 1997; 66: 1264-76.

  37. Linn T, Santosa B, Gronemeyer D, Aygen S, Scholz N,Busch M, Bretzel R.G. Effects of long term dietary proteinintake on glucose metabolism in humans. Diabetologia 2000; 43: 1257-65.

  38. Linn T, Geyer R, Prassek S, Laube H. Effect of dietaryprotein intake on insulin secretion and glucose metabolismin insulin-dependent diabetes mellitus. J Clin EndocrinolMetab 1996; 81: 3938-3943.

  39. Parker B, Noakes M, Luscombe N, Clifton P. Effect of ahigh-protein, high-monounsaturated fat weight loss diet onglycemic control and lipid levels in type 2 diabetes.Diabetes Care 2002; 25: 425-30.

  40. Patti,M, Brambilla E, Luzi L, Landaker E, Khan C.Bidirectional-modulation of insulin action of amino acids.J Clin Invest 1998;101:1519-29.

  41. Pi-Sunyer X Glycemic index and disease. Am J Clin Nutr 2002; 76: 290S-8S.

  42. Phinney S, Bistrian B, Wolfe R, Blackburn G. The humanmetabolic response to chronic ketosis without caloricrestriction: physical and biochemical adaptation.Metabolism 1983; 32: 757-768.Low carbohydrate diets

  43. Rosetti L, Rothman D, De Fonzo R, Schulman G. Theeffect of dietary protein on in vivo insulin action and liverglycogen repletion. Am J Physiol 1989; 257: E212-E219.

  44. Anderson JW, Konz EC. Obesity and diseasemanagement: effects of weight loss on comorbid condi-tions.Obes Res 2001; 9: 326S-334S.

  45. Dattilo AM, Kris-Etherton PM. Effects of weightreduction on blood lipids and lipoproteins: a metaanalysis. Am J Clin Nutr 1992; 56: 320-8.

  46. Lichenstein AH, Van Horn L. AHA Science Advisory.Very low fat diets. Circulation 1998; 98: 935-9.

  47. Parks EJ, Hellerstein MK. Carbohydrate-inducedhypertriacylglycerolemia: historical perspective andreview of biological mechanisms. Am J Clin Nutr 2000;71: 421-33

  48. Russell R, Taegtmeyer H. Pyruvate carboxylation preventsthe decline in contractile function of rat hearts oxidizingacetoacetate. Am J Physiol 1991; 257; E212-9.

  49. Best T, Franz D, Gilbert D, Nelson D, Epstein M. Cardiaccomplications in pediatric patients on the ketogenic diet. JNeurology 2000;54:2328-30.

  50. Fisler J. Cardiac effects of starvation and semi-starvationdiets: safety and mechanisms of action. Am J Clin Nutr 1992;56:230S-4S.

  51. Sours H, Frattali V, Brand C, Feldman R, Forbes A,Swanson R, Paris A. Sudden death associated with verylow calorie weight reduction regimens. Am J Clin Nutr 1981;34:453-61.

  52. Soloff L. Arrhythmias following infusions of fatty acids.Am Heart J 1970;80:671-5.

  53. Wadden T, Stunkard A, Brownell K. Very-low-caloriediets. Ann Intern Med 1983;99:675-84.

  54. Moss L. Caution: Very-low-calorie Diets can be Deadly.Ann Intern Med 1985; 102: 121-3.

  55. Baird I. Low-calorie-formula diets - are they safe? Int JObes 1981;5 (3): 249-56.

  56. Pietrobelli A, Rothacker D, Gallagher D, Heymsfield S.Electrocardiographic QTC interval: short-term weight losseffects. Int J Obes Relat Metab Disord 1997; 21 (2): 1-10.

  57. De Fonzo R, Soman V, Sherwin R, Hendler R, Felig P.Insulin binding to monocytes and insulin action in humanobesity, starvation and refeeding. J Clin Invest 1978; 62:204-213.

  58. Newman W, Brodows R. Insulin action during fasting andrefeeding in rats determined by euglycemic clamp. Am JPhysiol 1983; 249: E514-E518.

  59. Heaney RP. Nutrition and risk for osteoporosis. In:Osteoporosis. Marcus R, Feldman D, Kelsey J, eds. SanDiego: Academic Press, 1996; 483-505.

  60. Heaney R. The role of calcium in prevention andtreatment of osteoporosis. Phys Sports Med 1987;15:83-8.

  61. Osteoporosis Prevention, Diagnosis, and Therapy. NIHConsensus Statement 2000 March 27-29; 17(1): 1-36.

  62. Barzel US, Massey LK. Excess dietary protein canadversely affect bone. J Nutrition 1998;128:1051-3.

  63. Breslau NA, Brinkley K, Hill KD, Pak CY. Relationshipof animal protein-rich diet to kidney stone formation andcalcium metabolism. J Clin Endocrinol Metab 1988; 66:140-6.

  64. Lemann Jr J. Relationship between urinary calcium andnet acid excretion as determined by dietary protein andpotassium: a review. Nephronology 1999; 81 Suppl 1: 18-25.

  65. Wang X, Zhao X. The effect of dietary sulphur–containingamino acids on calcium excretion. Adv Exp Med Biol 1998; 442: 495-9.

  66. Kaneko K, Masaki U, Aikyo M, Yabki K, Haga A,Matoba C, Sasaki H, Koike G. Urinary calcium andcalcium balance in young women affected by high proteindiet of soy protein isolate and adding sulphur containingamino acids and/or potassium. J Nutr Sci Vitaminol 1990;36: 105-116.

  67. Promislow JH, Goodman-Gruen D, Slymen DJ, Barrett-ConnorE. Protein consumption and bone mineral densityin the elderly. Am J Epidem 2002; 155:636-44.

  68. Ball D, Maughan R. Blood and urine acid base status ofpremenopausal omnivorous and vegetarian women. Br JNutr 1997; 78:683-8.

  69. World Cancer Research Fund/American Institute forCancer Research. Food, nutrition and the prevention ofcancer: a global perspective. American Institute forCancer Research, Washington DC, 1997.

  70. Cummings JH, Bingham S. Diet and the prevention ofcancer. Brit Med J 1998; 317: 1636-40.

  71. Bingham S. Meat, starch and non starch polysaccharidesand large bowel cancer. Am J Clin Nutr 1988; 48: 762-7.

  72. Sandhu M, White I., McPherson K. Systematic review ofthe prospective cohort studies on meat consumption andcolorectal cancer risk: a meta-analytical approach. CancerEpidemiol Biomarkers Prev 2001;10: 439-46.

  73. Norat T, Riboli E. Meat consumption and colorectalcancer: a review of epidemiologic evidence. Nutr Rev 2001; 59: 37-47.

  74. Truswell A. Meat consumption and cancer of the largebowel. Eur J Clin Nutr 2002; 56 Suppl 1: S19-S24.

  75. Hill M. Meat cancer and dietary advice to the public. Eur JClin Nutr 2002; 56 Suppl 1: S36-S41.

  76. McIntosh G. Cereal foods, fibre and the prevention ofcancers. Aust J Nutr Diet 2001; 58: S34-S48.

  77. Xinhua C, Nayyar I, Guenther B. The effects of free fattyacids on gluconeogenesis and glycogenolysis in normalsubjects. J Clin Invest 1999; 103: 365-372.

  78. Brosnan JT. Comments on metabolic needs for glucoseand the role of gluconeogenesis. Eur J Clin Nutr 1999; 53Suppl 1: S107-11.

  79. Phinney S, Bistrian B, Wolfe R, Blackburn G. The humanmetabolic Response to chronic ketosis without caloricrestriction: physical and biochemical adaptation.Metabolism 1983; 32: 757-764.

  80. Hellerstein M. De novo lipogenesis in humans: metabolicand regulatory aspects. Eur J Clin Nutr 1999; 53 Suppl 1:S53-65.

  81. Schenker S. Dieting crazes. British Nutrition Foundation,Nutrition Bulletin 2001; 26: 117 -9.