Dear Professor Robinson,
Paper : Gestational IGT does not increase perinatal mortality in a developing country with high prevalence of glucose intolerance: an observational study
Thank you for suggesting you would be interested in this paper in short format. We enclose three copies of the revised version of this paper accordingly. We have considered the comments of the referee and statistician, and also include the other information requested, as follows:
1.Title of the paper
We agree with the comments about the title. We feel that as it now stands, it reflects the main message of the shortened paper, the poor outcome in pregestational diabetes in developing countries being well known.
2. Minimum criteria for gestational diabetes
These could be either two fasting plasma glucose or two random plasma glucose values in the diagnostic range or one of each. This is in line with WHO recommendations.
3. Perinatal mortality in North West and North East England (Page 12 of original paper)
This should have read as less than 10/1000 instead of less than 10 %. Now deleted.
4. Detection of IGT in pregnancy as risk factor for future diabetes
We agree that this is an important issue which merits further discussion in its own right, with regard to its relevance in primary prevention of diabetes in developing countries with high Type 2 diabetes prevalence, in the context of more urgent priorities in health care, limitations on health expenditure, and human rights issues which could be raised on grounds of gender and reproductive status.
Our shortened paper now concentrates on the more immediate perinatal implications, perinatal mortality and birth weight being among the most important current maternal and infant health problems in Mauritius, as in many developing countries.
5.Numbers of diabetic pregnancies vs numbers of women
We are not sure as to the reasons for the statistician’s suggestion that only first pregnancies for each woman in the study period should be included. Background population statistics for comparison are also based on numbers of deliveries rather than numbers of women, and may well include several pregnancies in individual women. Most existing diabetic pregnancy literature is based on numbers of pregnancies rather than numbers of women, and do not exclude subsequent pregnancies in the same women. We have retained consistency with the background data and current literature.
6.Background population statistics
These would most certainly be based on all deliveries, including those in women with diabetes or IGT. We agree that ideally they should not, but such population statistics excluding women with particular conditions are not available anywhere. The island perinatal mortality statistics were derived from over 60,000 births in 3 years and the background hospital birth weight and caesarean section data from around 4500 annual deliveries; any effect on the statistical analyses should be small.
7.Other statistical comments
No power statement is given, no pre-study calculation of required sample size having been made for this observational study recruiting all known patients in the study period. For pair-wise comparisons between the three groups, no multiple comparison tests were made as regards continuous data analysed by one way analysis of variance. Neither of these problems would seem to detract from the clear and consistent message the data give. We have not effected the recommended change for mean and SD notation as this would make the Table difficult to read.
8.Further details on percentages in text and tables
Please see attached tables (a) - (d)
I trust that the above satisfactorily answers the various issues raised by the reviewers and BMJ Editorial staff.
I shall be forwarding the form regarding competing interests as soon as this is returned duly signed by the authors in Mauritius.
Yours sincerely,
Shenaz Ramtoola MRCP MD
Consultant Physician
Royal Infirmary
Infirmary Road
Blackburn BB2 3LR
| Years of diabetes | n (%) |
| < 5 | 87 (79) |
| 5-9 | 16 (15) |
| 10+ | 7 (6) |
| Total | 110 (100) |
Table(b) Gestational age at initial diagnosis in women with gestational diabetes/IGT
| Gestational age (weeks) | Gestational diabetes | Gestational IGT | All |
| < 14 | 6 (7) | 4 (4) | 10 (5) |
| ³ 14, < 26 | 29 (34) | 15 (15) | 44 (24) |
| ³ 26, < 37 | 47 (55) | 73 (74) | 120 (65) |
| ³ 37 | 4 (5) | 6 (6) | 10 (5) |
| Total | 86 (100) | 98 (100) | 184 (100) |
n (%)
Table(c) Reasons for plasma glucose measurements in gestational onset
diabetes or IGT
| Reason for test | Gestational diabetes | Gestational IGT | All |
| Glycosuria | 68 (79) | 73 (75) | 141 (77) |
| "At risk" | 7 (8) | 13 (13) | 20 (11) |
| Screening | 4 (5) | 6 (6) | 10 (5) |
| Symptoms | 3 (3) | 0 (0) | 3 (2) |
| Other/Unknown | 4 (4) | 6 (6) | 10 (5) |
| Total | 86 (100) | 98 (100) | 184 (100) |
n (%)
Table(d) Outcome of pregnancy according to diagnostic category of glucose intolerance
Stillbirth 6/89 (67) 7/86 (81) 1/92 (11) 901/63442 (14)
Early neonatal mortality 5/83 (60) 3/79 (38) 1/91 (11) 765/62541 (12)
Macrosomia (³ 4000 g) 7/89 (8) 14/86 (16) 6/92 (7) 147/4440 (3)
Prematurity (<37 wk) 33/89 (37) 19/86 (22) 10/92(11) NA
Caesarean section 51/88 (58) 49/86 (57) 40/92 (43) 2198/13449 (16)
Hypoglycaemia 17/80 (21) 11/78 (14) 4/90 (4) NA
Hyperbilirubinaemia 28/80 (35) 30/78 (39) 19/90 (21) NA
n (rate per thousand) for mortality, otherwise n (%)
hypoglycaemia and hyperbilirubinaemia data not available in 5 infants, and mode of delivery not recorded in 1 patient (pregestational). Background statistics over 3 years except birth weight available for one year only.