R. M. Rizvi ( Department of Obstetrics and Gynaecology, The Aga Khan University Hospital and Baqai Medical University*, Karachi. )
P. Baillie ( Department of Obstetrics and Gynaecology, The Aga Khan University Hospital and Baqai Medical University*, Karachi. )
June 2003, Volume 53, Issue 6
Original Article
Introduction
Pakistan is the seventh most populous country in the world and has the highest rate of growth among these countries. The maternal mortality rate is also very high; each year 340 /100,000 women die in the process of child birth, and the major causes identified are sepsis, haemorrhage and infection.
This communication describes the outcome of seriously ill pregnant or recently pregnant women admitted to an Intensive Care Unit (ICU) in a developing country when management was altered in accordance with ICU principles, modified for the pregnant state and aimed at easy application to developing countries.
Subjects and Methods
All women who were pregnant or recently pregnant within one month and considered ill enough to be admitted to the ICU of Baqai Medical University were studied retrospectively over a 2-year period (January 1997 to December 1998) and prospectively for 2 ½ years (Jan 1999 to June 2001) after changes in management were introduced.
During the retrospective study, 359 patients were admitted as emergencies to obstetrics and gynecological wards, 49 were admitted to the ICU (only 2 booked hospital patients, the rest being outsourced). Of these 49 patients, 29 (58%) died. During the prospective study there were 436 emergency admissions with 22 admissions to the ICU. Of these 4 (18%) died.
A detailed description of management of both retrospective and prospective patients is as follows:
1. In prospective group an early diagnosis of cardiovascular compromise by means the systemic inflammatory response syndrome (SIRS)3 was made in addition to refill time of plasma blanching as a test of capillary flow.4 The diagnosis of SIRS depends on 2 or more of the following features: blood pressure 90/60mmHg, pulse rate 100beats/min, respiratory rate 22 breaths/min, temperature 38.50C or 36.00C white cell count 15,000/cumm or 4000.
These have been modified from normal ICU values to allow for pregnancy changes. Progression of the disease process results in the multiple organ dysfunction syndrome (MODS), which is defined according to the number of organs involved, and dysfunction as follows:4 Respiratory pO2/FiO2 <225 or respiratory rate >23 breaths/minute, serum creatinine >100umol/litre, platelets <100,000 per sq mm, pressure adjusted pulse rate >15, bilirubin >20umol/litre, Glasgow score <14.
The skin and gastrointestinal tract have been added recently. Organ involvement generally carries a mortality of 15-20% per organ involved. All patients had >4 organs involved with an expected mortality of up to 60%.5
2. In retrospective study 5% dextrose water as 3 liters/24 hours was used and no CVP line was passed. While in prospective study where cardiovascular compromise was detected, an immediate infusion6 of 3 liters of Ringers lactate was carried out over 1-2 hours and continued until the condition settled or adequate volume stabilization was achieved by the attainment of a Central Venous Pressure (CVP) of 10 to 15cm H2O. This was usually measured indirectly in an oblique light assisted by supraclavicular compression of the internal and external jugular veins, the latter being easily approached. Cardiac compromise from myocardial dysfunction or fluid overload was prevented by keeping the Pressure-adjusted Pulse Rate (PAR) below 20 in acute circumstances and 15 beats in less acute conditions. The PAR is calculated as follows.4
| Pulse rate x CVP PAR = ___________________________ Lowest mean blood pressure Lowest mean Blood Pressure (BP) in 4 hours if sub-acute, and 20 minutes in acute shock. Young patients tolerated 15 Liters (L) or even 20 L well in contrast to older patients with intrinsic heart disease where 10 L may have serious consequences. In severely ill patients where the blood pressure was not restored despite adequate volume replacement, vasopressin was used as the drug of choice7 rather than dopamine or dobutamine used in retrospective patients. This was only required once in sharp contrast to the retrospective study where dopamine was freely used with lesser infusion rates over the first 24 hours based on urinary output, often with diuretics, which were avoided in the prospective study. Oxygen4 was used by facial mask at high flow rate (>5 liters/minute) if the respiratory rate was more than 22 per minute. Dextrose water and saline were not used. The only colloid was 1L 6% hydroxyethyl starch when increasing edema supervened despite a need for further infusions.8 Fresh frozen plasma and blood were used very sparingly. Blood components, especially platelets, were used in consultation with the haematologist. The prospective and retrospective studies did not alter obstetrics and gynaecological management with surgical interventions carried out after stabilization unless in extreme situations. All patients were seen within 30 minutes of admissions. Antibiotics were discontinued after 72 hours in the prospective study unless a clear indication was present for continuation. Postmortems were uniformly refused. The 4 deaths in the prospective study were subjected to confidential inquiry. Statistical analysis was performed using the Epinfo 2000 computer statistics programme. To compare proportions, the c2 test and Fisher's exact test (when expected numbers will be less than five) were used. A sample size of 45 in retrospective group against 15 patients in prospective period achieved 80% power (1-b) was required to detect decrease in maternal mortality of 40% with 5% level of significance (a). |
Results
Table 1. Major interventional differences between retrospective and prospective studies.
| Differences | Prospective | Respective(excluding SIRS) |
|
In 11 out of 21 cases, clinically diagnosed infections were confirmed on blood cultures. This is consistent with the literature5 and has implications for causation and antibiotic use. No statistical significant difference was found in the probability of culture being positive in patients with sepsis and septic shock between prospective and retrospective time period RR 1.22 (95 % CI 0.47-3.16, p 0.06) Table 2.
Table 2. Bacteriological culture reports.
| Prospective culture position | Retros pective culture position | RR | CI | P value |
| ||||||||||||||||||
| Table 3. Survival status by initiating condition. |
| Initiation condition | Prospective n=22 Survivors | Retorspective n=49 Survivors | RR | CI 95% | P Value |
|
| Total survivors | 18(81.8%) | 20(40.8%) | 2.00 | 1.36-2.96 | 0.001 |
During prospective study period, significantly higher proportions of patients survived 81.8% vs 40.8% (p 0.001) (Table 3). Survival was significantly higher in the sub group of patients presenting with sepsis 87.5% vs 30.7% (p 0.024). Although the number of emergency admissions were similar, far fewer patients were admitted to ICU during the prospective study period i.e., 22/436 (5.0%) vs 49/359 (13.6%) (p <0.0001).
Table 4 summarizes survival status by disease process. Significantly higher survival rates were observed in patients presenting in prospective time period with SIRS progressing to MODS and also for ones who presented primarily with MODS.
Discussion
Our study in contrast, showed that the common disease process (SIRS progressing to MODS, and MODS on admission) was a better predictor of maternal outcome. Furthermore increased attention to the disease process in a simple manner using ICU principles of early detection and adequate fluid infusion, modified for ease of application in the developing world resulted in a major decline in admissions to ICU and also maternal mortality. Obstetric and gynecological management remained constant.
| Table 4. Survival status by disease process. |
| Diseases | Prospective Survivors n=22 | Retorspective Survivors n=49 |
|
The central ICU principle is that a wide variety of conditions10 (trauma, burns, infections, pre-eclampsia, hemorrhage, pancreatitis, coronary thrombosis and many others) produce the same metabolic response of SIRS followed by MODS if the process is life threatening or inadequately resuscitated.10 Reperfusion may produce further problems11, although not seen in our patients.
The mechanism is a disequilibrium between the nonspecific inflammatory cytokine process and immune defence mechanism. The process of SIRS MODS and the later onset of Compensatory Autoimmune Response Syndrome12 (CARS) follows a predictable clinical course over several days or even weeks unless compressed by severity of the disease or inadequate management and resuscitation resulting in early death. Initially ebb (hypotension) is followed by flow (hyperdynamic phase.)10 When the somatic insult is severe enough to escape local defence mechanisms, SIRS is produced by inflammatory cytokines such as TNFa and serves as an early warning symptom that the process is no longer localized although in general the defence mechanisms are usually adequate.13 Nevertheless early warning symptoms are critical to outcome as well as adequate early management. The use of SIRS rather than the traditional vital signs was an important part of the strategy for reducing maternal mortality and limiting progression that required ICU admissions.
Similar to other studies, we found proven infection in the minority of apparently clinically infected patients.5 The temperature may well be due to inflammatory cytokines such as TNFa. In addition it has been suggested that the majority of continuing fevers may well be due to polymorphisms of TNFa and other inflammatory cytokines i.e. the response of the patient.13
The other noteworthy feature of our study was the high volumes of fluids infused.14 Ten liters of fluids are lost to the circulation following a major operation and this is not corrected for up to 2 weeks in older patients14 and postpartum loss of 400 ml of blood requires more than 2 liters of crystalloids to equilibrate the circulatory volume.14 It has been questioned whether irreversible shock ever occurs if enough fluid is infused-upto 530 ml/kg have been used in experimental animals with 10 times the estimated blood volume being infused.
Fluid overload occurred in only 2 patients in our study emphasizing the general health of young women and perhaps the use of JVP and PAR in acute circumstances to detect a myocardial component. Of most importance and curiously deficient in obstetric literature, is the occurrence of hyponatremia as a contributing factor to fluid overload.15 Both our patients were associated with the infusion of high doses of oxytocin in 5% dextrose water. The death from eclampsia was almost certainly due to hyponatremia and not eclampsia. Acute hyponatremia, the commonest cause of death postoperatively15 is lethal in young women and is usually unrecognized. This also emphasizes that there is no place for hypotonic fluids such as 5% dextrose water in acutely ill women.16 Iatrogenic hyponatremia is inexcusable.
We used Hartmann's solution or lactated ringers only because of cost considerations, thereby not entering the colloid/crystalloid debate. The only colloid used was 6% hydroxyethyl starch8,16 when capillary damage was evidenced by increasing peripheral edema without volume repletion. This had little apparent effect on our outcome and it should be noted that even normal saline produces a hyperchloraemic acidosis which is not helpful in an acutely ill patient.
Albumin has been implicated in a 6% increased mortality so the widespread use of fresh frozen plasma must be questioned, particularly as the usual pathology of DIC is decreased platelets and very seldom and transiently hypofibrinogenemia. The bedside detection of clothing defects in acutely ill women, remains a problem, although advances in thromboelastography will probably solve17 this. At present we use hand incubated whole blood clotting time with laboratory Fibrinogen Degradation Products (FDP's) and platelets as an adjunct. The involvement of the liver in these patients makes Activated Partial Thromboplastin Time (APTT) and International Normalization Ratio (INR) of limited use in acute management.
It should be noted that the obstetric dictum "if she has lost blood, replace blood" is questionable as full transfusion is associated with an increased mortality17 in actually ill patients, particularly in younger patients. The reasons for this are probably multiple. Blood is certainly pyrogenic and proinflammatory, which may well be deleterious in SIRS and MODS. Transfusion can produce respiratory distress syndrome and may not be available for oxygen exchange because of 2-3 DPG degradation particularly in banked blood. For these reasons, most guidelines state that above 10gm% (or even 8gm%) transfusion is seldom necessary. Below 6gm% it is almost always required and between 6 and 10gm% the clinical condition dictates usage i.e. conservative use of blood.18
The overriding clinical lesson of our study was that in acute circumstances simple BP and intake output calculations of fluid requirements are often disastrous resulting in deterioration as they do not take into account the circulatory requirements, particularly for adequate splanchnic perfusion.
During the immunological changes of pregnancy however, a less effective defence mechanism is present with decreased cell mediated immunity and compensatory increase in innate immunity resulting in the well known increased likelihood of septic shock.19
Women also have less effective defence mechanisms against hypotension than men and pregnancy aggravates this with peripheral vasodilatation from nitric oxide and prostanoids20, alterations in baro-receptors sensitivity and decreased sensitivity to catecholamines despite a compensatory increase in circulating catecholamines in later pregnancy21 together with volume and respiratory changes. These account for the altered definition of SIRS in pregnancy as well as our decision to use vasopressin for volume replete hypotension rather than dopamine22, which is still surprisingly widely used.
The final issue is specific treatment of SIRS and MODS. The essential management of SIRS and MODS requires early recognition of vascular compromise in a wide variety of obstetric and gynaecological conditions as well as adequate volume repletion and maintenance of circulating volume with appropriate fluids as soon as possible if lethal MODS is to be avoided.
In summary, we have shown a marked reduction in seriously ill pregnant women as well as maternal deaths by early detection of vascular compromise and adequate volume repletion. The methods were chosen to be easily applicable clinically in the developing world and it is strongly suggested that attention to the generalized disease process of SIRS, MODS and CARS is at least as important as the initiating condition and should be included in all protocols, particularly in the developing world where maternal mortality rates remain obstinately high.
Refrences
2. Task Force of the American College of Critical Care, Medicine. Practice Parameters for hemodynamic support of sepsis in adult patients in sepsis. Crit Care Med1999;27:639-60.
3. Malone DL, Kuhls D, Napolitanol M, et al. Back to basics: validation of the admission SIRS score in predicting outcome in trauma. J Trauma 2001;51: 458-63.
4. Evans TW, Smithies M. ABC of intensive care: organs dysfunction. Br Med. J 1999;318:1606-9.
5. Wheele, AP, Bernard GR. Treating patients with severe Sepsis. N Engl J Med 1999;340:207-14.
6. Ansus DC, Linde WT, Lidickerj, et al. Epedimiology of severe sepsis in the US: analysis of medicine outcome and associated cost of care. Crit Care Med 2001;29:1303-10.
7. Hind CJ, Watson D. ABC of intensive care circulatory support. Br Med J 1999;318:1749-52.
8. Nolan J. Fluid replacement. B. Med Bull 1999;55:821-43.
9. Lapinsky SE., Kruczyasi K, Balter MS, et al. Obstetrics Patients in ICU. Crit Care Med 1994;22:840-2. 10. Foex BA. Systemic responses to trauma. Br Med Bull 1999;55:726-43.
11. Collard .D, Gelman S. Pathophysiology, clinical manifestation and prevention of ischaemia-reperfusion injury. Anaesthesia 2001;94:1133-8.
12. Mannick JA, Rodrick ML, Lederer JA., et al. The immunological response to injury. J Am Coll Surg 2001;193:237-44.
13. Mira JP, Cariou A, Grau F, et al. Association of TNF2, a TNF a promotion polymorphism with septic shock susceptibility and mortality. JAMA 1999; 282:561-8.
14. Healey MA, Samphire J, Hoyt DB, et al. Irreversible Shock is not irreversible. A new model of massive hemorrhage and resuscitation. J Trauma 2001;50:826-34.
15. Fraser CL, Arieff AI. Epidemiology, Pathology, Physiology and management of hyponataemic encephalopathy. Ann Intern Med 1997;102:67-77.
16. Wilkes NJ, Woolf NR, Mutch M, et al. The effects of balanced versus saline based hetastarch and crystalloid solutions on acid-base and electrolyte status and gastric mucosal perfusion in elderly surgical patients. Anesth Analg 2001;93;811-6.
17. Brazil EV, Coats TJ. Sonoclot coagulation analysis of in vitro haemodilution with resuscitation solutions. J R Soc Med 2000;93:507-10.
18. Hebert PC, Wells G, Blajchman MA, et al. A multicentre randonmized clinical trial of transfusion requirements in critical care. N Engl J Med 1999; 340:409-17.
19. Vizi ES, Szelenyi J, Papp Z, et al. Enhanced TNFa specific and decreased IL10 specific immune responses to LPS during the third trimester of pregnancy in mice. J Endocrinol 2001;171:355-61.
20. Greer IA, Walker JJ, McLaren M, et al. Immunoreactive prostacyclin and thromboxane metabolites in normal pregnancy. Br J Obstet Gynaecol 1985; 92:581-5.
21. Greenwood JP, Scott EM, Stoker JB, et al. Sympathetic neural mechanisms in normal and hypertensive pregnancy in humans. Circulation 2001;104:2200-4.
22. Forrest P. Vasopressin and shock. Anaesth Intensive Care 2001; 29: 463-72.
Journal of the Pakistan Medical Association has agreed to receive and publish manuscripts in accordance with the principles of the following committees:
