weight, with an operative mortality of obese people? Lancet 2001; 357: 1357–59. causes of death, and show a greater
0·5%.2 2 Scopinaro N, Adami GF, Marinari GM, decrease for causes that previously
Questioning whether surgeons et al. Biliopancreatic diversion. World J Surg showed a larger increase. Use of a
1998; 22: 936–46.
should treat type 2 diabetes in severely correlation coefficient would give such a
3 Pories WJ, Swanson MS, MacDonald KG,
obese patients is unjustified at present. et al. Who would have thought it? An result when applied to chaotic changes
They should and they are treating such operation proves to be the most effective in any semistable system.
patients. The 1991 National Institutes therapy for adult-onset diabetes mellitus. Shkolnikov and colleagues ack-
for Health consensus conference Ann Surg 1995; 222: 339–50. nowledge that the quality of coding of
4 Mingrone G, De Gaetano A, Greco AV,
already included, among indications for et al. Reversibility of insulin resistance in
deaths in Russia is poor, as are data on
obesity surgery, type 2 diabetes (body- obese diabetic patients: role of plasma lipids. consumption of alcohol. We agree that
mass index >35 kg/m2). In series with Diabetologia 1997; 40: 599–605. “changes in acute alcohol poisoning can
long-term follow-up, gastric bypass and 5 MacLean LD, Rhode BM, Sampalis J, be regarded as . . . estimation of
BPD achieve durable normal concen- Forse RA. Results of the surgical treatment changes in the frequency of excessive
of obesity. Am J Surg 1993; 165: 155–60.
trations of plasma glucose, plasma drinking”, but we do not know for
insulin, and glycosylated haemoglobin sure how reliable this estimation is.
in 80–100% of severely obese diabetic Mortality in Russia They postulate that some deaths from
patients.2,3 alcohol poisoning are registered
On the basis of the exceptional Sir—Vladimir Shkolnikov and col- as cardiovascular deaths, but this
control of diabetes achieved in patients leagues’ report (March 24, p 917)1 on hypothesis is unproven. Probably
with severe obesity, we should ask, changes of mortality in Russia is misclassification prevails in the other
instead, whether gastric bypass and thought provoking. The researchers direction.
BPD represent a cure for type 2 acknowledge the increase in longevity The alcohol explanation is so
diabetes in moderately obese (body- after 1994, which they ignored until attractive, that Shkolnikov and
mass index 30–35 kg/m2) or non-obese 1998,2 but persist in explaining the colleagues even connect to drinking the
patients. Two sources have suggested mortality changes by Russians’ devil decline in deaths from motor-vehicle
this effect is possible. Remission of type drinking. accidents. They wrongly state that in
2 diabetes has been noted in non-obese They explain the decrease in 1991–94 the number of cars on roads
individuals undergoing BPD for other mortality by reduction in alcohol lessened. Since 1990, the number of
indications,4 and glucose and insulin consumption. The alcohol hypothesis cars has risen steeply.
concentrations normalise within days of was not sufficiently supported in *V Vlassov, V Gafarov
gastric bypass or BPD, before weight Shkolnikov and colleagues’ earlier *Saratov Medical University, PO Box 1528,
loss occurs.3 The control of diabetes reports3 and it continues to be Saratov 410601, Russia; and Novosibirsk,
seems, therefore, to be a primary, unsupported in this study. The Institute for Therapy, Novosibirsk
specific, and independent effect of connection between mortality and
these operations and not secondary to social stresses was also seen in Russian 1 Shkolnikov V, McKee M, Leon DA.
treatment of obesity. participants of the MONICA project, Changes in life expectancy in Russia
Although the mechanism is but they believe that although alcohol is in the mid-1990s. Lancet 2001; 357:
unknown, bypass of duodenum and an indicator of stress, there is no proof
2 Leon DA, Shkolnikov VM. Social stress and
proximal jejunum, a common feature of that increase of mortality is caused by the Russian mortality crisis. JAMA 1998;
both gastric bypass and BPD, probably alcohol consumption.4 279: 790–91.
contrasts a hormonal or neural signal When total increase in total mortality 3 Vlassov VV. The role of alcohol and social
originating from the gut in response to in 1981–94 was 718 per 100 000, the stress in Russia’s mortality rate: recent
mortality trend in Russia. JAMA 1999; 281:
the passage of food, which leads to the input of mortality from alcohol 321–22.
impaired action, secretion, or both of poisoning was only 56 per 100 000. 4 Gafarov VV, Gagulin IV. Population study
insulin. This statistic must discourage the of socio-psychological risk factors of
Pinkney and colleagues ask also hypothesis that alcohol is the main coronary heart disease in male population of
which is the most appropriate operation cause of mortality changes. The real Novosibirsk. Ter Arkh 2000; 4: 40–43.
for obese patients with type 2 diabetes, statistic is even worse: cited numbers
but give no clear answer. Gastroplasties are for ages 15–74 years, when alcohol- Sir—Vladimir Shkolnikov and col-
improve glucose metabolism but there related mortality is relatively increased. leagues1 suggest that small variations in
is a lack of evidence that they can Shkolnikov and colleagues go further: alcohol consumption per person in
induce stable remissions of diabetes. they first present data for age group Russia during 1984–98 caused huge
Eight different randomised controlled 15–24 years, then they accumulate all fluctuations in life expectancy and
trials have compared vertical-banded external causes of death (about 80% mortality rates. This phenomenon
gastroplasty with gastric bypass and of mortality in this age) and connect needs to be clarified.
workers have shown a clear long-term these deaths to increased alcohol Some other European countries have
weight loss to be better with gastric consumption. For all ages they interpret had similar or even greater changes in
bypass.5 We believe morbidly obese cirrhosis of the liver as alcohol-caused alcohol consumption per person
patients with type 2 diabetes would disease with no reservations. Any cause without such variations in mortality.
benefit more from gastric bypass or of death is connected to alcohol by In Europe, each 1 L variation in
BPD than from gastroplasty. selective citation of studies supporting consumption changes mortality by
*Francesco Rubino, Michel Gagner, this connection. about 1%.2 However, the decrease of
Jacques Marescaux Trying to prove their alcohol total alcohol consumption by 3 L
*IRCAD-EITS, Louis Pasteur University, 67091 hypothesis, Shkolnikov and colleagues during an anti-alcohol campaign in
Strasbourg, France; and Mount Sinai Medical use a procedure that might lead to 1985–87 was followed by a rapid
Center New York, NY, USA biased conclusions. They calculate the decline of total mortality in Russia of
correlation coefficient of the increase 12%. The return to the initial level of
1 Pinkney JH, Sjöström CD, Gale EAM. compared with the subsequent decrease alcohol consumption in 1992–93 (at
Should surgeons treat diabetes in severely of the mortality for the number of the beginning of market reforms)
THE LANCET • Vol 358 • August 25, 2001 669
For personal use. Only reproduce with permission from The Lancet Publishing Group.