CHAPTER XV
Postoperative Complications: Prophylaxis and Therapy
James C. Drye, M. D., and W. Philip Giddings, M. D.
The records of the 2d Auxiliary Surgical Group are not complete in respect to postoperative complications and are particularly fragmentary in respect to less severe and nonfatal types. For these deficiencies, there are two explanations. The first is, as already pointed out, that the circumstances of war did not permit the keeping of clinical records which were complete and accurate in all details. The second is that many postoperative processes which in civilian practice would be regarded as complications were so frequent in battle wounds that they came to be regarded as almost implicit in certain injuries and after certain operations. Such complications were seldom made a matter of record. That is why, in this series, it is impossible to make any accurate determination of mild incisional sepsis, ileus of brief duration, pulmonary complications of minor severity, and similar processes.
Early in the war, it was found that premature evacuation of casualties with abdominal injuries increased the incidence of wound dehiscence, ileus, peritonitis, and other major postoperative complications. The trauma and discomfort of ambulance transport over rough roads, as well as the loss of continuity of treatment in a critical stage of convalescence, made these results almost inevitable. It therefore became the policy that, whenever the tactical situation permitted, patients with abdominal wounds should not be evacuated earlier than the eighth day after operation (p. 85). In practice, evacuation was delayed for 14 days in a great many cases, and, if the injuries were unusually severe, it was sometimes delayed for as long as 30 days. The tactical situation frequently prevented complete adherence to this policy, but the majority of the patients in this series were nevertheless held in forward hospitals from 8 to 14 days. It was the emphatic opinion of the surgeons of the group that this policy, quite as much as any other consideration, produced the low morbidity and case fatality rates generally observed in casualties with abdominal injuries in World War II.
Only 3,090 of the 3,154 records used in this analysis were sufficiently complete to permit an analysis of the presence or absence of postoperative complications. All of the complications discussed are also discussed under special headings in other sections of this report. In spite of the inevitable repetition, however, it was thought worthwhile to bring them together at this point and comment upon them briefly.
204
PULMONARY COMPLICATIONS
In spite of the known frequency of postoperative pulmonary complications,their recorded occurrence is less accurate than that of other complications inthis series (table 31). In addition to the general reasons already listed, stillother reasons can be advanced to explain this situation: An accurate diagnosisof pulmonary conditions was likely to be difficult under field conditions. Eventhough an adequate physical examination was sometimes impossible, the diagnosisusually had to be based entirely on physical findings. Postoperativeroentgenograms of the chest were seldom made. When they were, they were usuallyunsatisfactory because of the limited technical facilities in forward hospitals.In many other instances, examination was precluded by the precarious state ofthe patient or by the presence of heavy plaster casts.
Sputum examinations were seldom made, because bacteriologic facilities werenot easily available. For these various reasons, the statistics tabulated areadmittedly deficient. They are, however, in accord with mature, clinicalimpressions, and the trends which they indicate can be accepted as accurate.
Pulmonary complications were frequent after operation in patients with abdominalinjuries, as might have been expected, because conditions in the theater favoredthe development of such complications. A large proportion of the casualtiesoccurred during the period of the year which is cold and wet in Italy (p. 216). Troops fought in mud and were almost constantly exposed toheavy rainfall or snow from October through March. As a result, more casualtieswere received with established respiratory infections during this period thanduring the summer months, the common cold following the same seasonal trend inItaly as it does in the temperate zones of the United States. Furthermore, manypatients passed the critical hours immediately before and after operation indamp tents (fig. 26), inenvironmental temperatures of 60? or 50? F., or lower. All of thesecircumstances combined to cause a high incidence of infectious pulmonarycomplications, particularly during the winter months.
The recorded case fatality rate, almost 40 percent(table 31), is further proof that only the most severe pulmonary complicationswere listed. It was the general impression that their severity, as well as theirfrequency, was greater during the winter than during the summer, and the casefatality rates for the two periods lent support to this opinion. Availablefigures, however (table 31), suggest that the presence of associated chestwounds did not appreciably affect either the frequency or the case fatality rateof the usual infectious pulmonary complications which occurred in abdominalcasualties.
There seems no doubt that most of the infectious complications wereatelectatic in character and that they arose on the basis of bronchitis or someother respiratory infection already present at the time of operation. In otherwords, they were those which might be expected to develop in a patient
205
FIGURE 26.-Conditions of surgery infield hospital in Italy in winter of 1944-45.
Resuscitation tent.
population of combat troops carrying a heavy prewounding load of endemicrespiratory infections.
The only two instances of lung abscess, both of which were fatal, were notassociated with thoracic injuries. Empyema, on the other hand, was associatedwith chest injuries in 25 ofthe 29 recordedcases, 11 of which were fatal. All 4 fatalities which occurred in the 7instances of bile empyema were associated with thoracoabdominal wounds.
The fact that hydropneumothorax was recorded only 91 times in the 3,090 abdominal injuries suggests thatnote was taken of only the more serious cases, in which repeated thoracenteseswere necessary. The incidence was undoubtedly much higher. In most instances, 2to 3 aspirationsof the pleural cavity were sufficient.
The actual incidence of both pulmonary edema and the so-called traumatic wetlung is known to be considerably greater than the recorded figures indicate(table 31). Pulmonaryedema was readily precipitated in severely wounded patients, especially in thosein shock, if intravenous fluids were given in too great quantities or toorapidly.
As was pointed out in the discussion of the multiplicity factor, theincidence of postoperative pulmonary complications tended to rise with theincrease in the number of viscera injured, the observation being, however, ofclinical rather than statistical significance (table 32). The explanation was simple: themore
206
TABLE 31.-Distribution ofrecorded pulmonary complications in 3,090 abdominal injuries
Complication | With thoracoabdominal or associated chest wounds (965)1 |
| Total | |||
| Deaths | Cases | Deaths | Cases | Deaths | |
Hydropneumothorax | 86 | 9 | 5 | 0 | 91 | 9 |
Atelectasis | 23 | 11 | 37 | 5 | 60 | 16 |
Bronchopneumonia | 12 | 6 | 39 | 19 | 51 | 25 |
Empyema | 25 | 9 | 4 | 2 | 29 | 11 |
Pulmonary embolism | 5 | 5 | 17 | 16 | 22 | 21 |
Pulmonary edema | 6 | 4 | 16 | 11 | 22 | 15 |
Bronchopleural fistula | 18 | 5 | 0 | 0 | 18 | 5 |
"Consolidation" | 6 | 4 | 8 | 4 | 14 | 8 |
Wet lung | 4 | 3 | 9 | 4 | 13 | 7 |
Aspiration of vomitus | 1 | 1 | 7 | 5 | 8 | 6 |
Lobar pneumonia | 2 | 2 | 5 | 5 | 7 | 7 |
Bile empyema | 6 | 4 | 1 | 0 | 7 | 4 |
Blast | 2 | 1 | 4 | 3 | 6 | 4 |
Bronchitis | 2 | 1 | 1 | 1 | 3 | 2 |
Lung abscess | 0 | 0 | 2 | 2 | 2 | 2 |
Other | 1 | 1 | 6 | 0 | 7 | 1 |
| 199 | 66 | 161 | 77 | 360 | 143 |
1Associated wounds include nonpenetrating trauma to the chest wall.
Item |
| Total cases | ||||
| Two | Three | Four | Five | ||
Cases | 1,348 | 1,014 | 350 | 96 | 23 | 2,831 |
Complication | 87 | 55 | 40 | 11 | 3 | 196 |
Incidence (percent) | 6.5 | 5.4 | 11.4 | 11.5 | 13.0 | 7.0 |
severely wounded patients were the ones most likely topresent stagnation of the tracheobronchial secretions, and atelectasis andpneumonia were therefore more likely to develop in them.
Prophylaxis and therapy-The prophylaxis and therapy ofpulmonary complications were both based on a few fundamental principles. Themost efficient prophylaxis consisted of expertly administered anesthesia, withpar-
207
ticular attention to a careful tracheobronchial toilet at theclose of the operation. Catheter aspiration or aspiration bronchoscopy was partof the established routine in patients with thoracoabdominal injuries andassociated chest wounds and was frequently carried out in patients with othertypes of injuries, just before they were taken off the operating table.Dressings were applied so as not to restrict the respiratory excursion.
After operation, the first essential was to maintain a cleartracheobronchial tree. For this, frequent changes of position and frequentcoughing were essential. In patients with multiple wounds, particularly those inlarge, bulky casts, changes of position were not always easy to accomplish. Firmmanual pressure was of great aid to patients endeavoring to cough, and manylearned to support their own incisions.
If the patients refused to cough or could not cough satisfactorily,tracheobronchial suction was again resorted to. If it was not effective,bronchoscopy was employed without delay. Oxygen therapy was employed accordingto the indications. All of these measures, which were standard practice,approximate the usual methods of prevention of postoperative complications incivilian practices.
Patients with wet lungs, who required repeated tracheal aspirations to clearthe airways effectively, were usually treated with tracheal catheter in situ. Itwas aspirated at regular intervals, and oxygen was usually administered in theinterim, sometimes under slight pressure.
Morphine was administered in as small doses and as infrequently as possible,to prevent depression of the respiration. The amount necessary could frequentlybe reduced by the use of nerve blocks for the relief of pain. Some surgeons ofthe group made it a practice to block the intercostal nerves from within, whilethey were exposed in thoracoabdominal wounds, or to perform nerve block beforethe patient was removed from the operating table.
Pulmonary edema could usually be prevented by care in the administration ofintravenous fluids. Hypodermoclysis was occasionally resorted to instead ofinfusion, but the method was not looked upon with favor because of thepossibility of introducing anaerobic organisms. In a few instances in whichpulmonary edema appeared soon after operation, venesection was done, withapparent benefit. Preparations of digitalis by the parenteral route also seemedof benefit. These measures were resorted to on empirical grounds. Atropine wasused according to the indications.
Although penicillin and sulfadiazine were given routinely aspart of the prevention and treatment of pulmonary complications, pneumoniasstill developed in patients who were adequately treated by these agents. It wasthought, although this was merely a clinical impression, that patients withpneumonic processes who did not respond to chemotherapy and antibiotic therapy,or who died under proper treatment, might have the so-called atypical variety ofpneumonia.
208
NONPULMONARY COMPLICATIONS
Shock.-Shock was the most important of all postoperative complications (p.119). Its correct incidence is undoubtedly greater than the figures suggest. Therole it played in the case fatality rate in these 3,090 casesis shown by the fact that 64 percent (472) of all the deaths (737) were attributable to it. It was the most frequent cause ofdeath in the first 48 hours afteroperation, and it was almost invariably fatal in the small number of patients inwhom it persisted after this time.
The essentials of management were blood replacement, oxygen by nasal catheteror the Boothby mask, and careful analgesic medication administered with due careto prevent further depression. These measures were instituted as soon as thepatient was seen and were continued after operation as long as necessary. Whenmorphine was indicated, the dose was kept small, and it was preferablyadministered by the intravenous route.
Peritonitis.-Peritonitis of some degree existed a priori inall patients with perforations of hollow viscera. Its frequency can be estimatedfrom the fact that injuries of this type were present in more than half of allcases in the series. The true ante mortem incidence of invasive bacterialperitonitis is not known. It existed in all degrees of virulence, sometimes inso mild a form that the diagnosis was merely academic and sometimes, at theother extreme, in such overwhelming contamination and infection that the patientdied within a few hours. The great majority of patients who presented the usual clinicalpicture of ileus, abdominal tenderness, and fever recovered uneventfully, but 91(12.3 percent) of all the deaths were due to this cause.The most effective measures in both prevention and therapy were decompression ofthe gastrointestinal tract and routine parenteral administration of sulfadiazineand penicillin.
Abdominal distention-Abdominal distention, whether as a manifestation ofperitonitis or from other causes, was uncommon in this series, undoubtedly asthe direct result of the routine practice of postoperative decompression of thegastrointestinal tract. Two- or three-bottle siphonage suction was used, by theWangensteen method; discarded 1-liter flasks originally used for intravenoussolutions were excellent for this purpose. The tube was usually left in placefrom 3 to 6 days.It was essential to leave it in situ until there was definite evidence of returnof intestinal tonus and motility; most instances of distention followed itspremature removal.
This simple method of preventing and controlling abdominaldistention and subsequent ileus was chiefly responsible for the low incidence ofa complication which, with its sequelae, had plagued military surgeons inearlier wars.
Intestinal obstruction-Mechanical intestinal obstruction was recorded in21 cases, in all but 1 instance in patients with wounds of the small intestine(p. 250). The figure is probablycorrect; the condition is too serious a complication to be overlooked.Operation was performed in only 3 patients,all of whom survived; there were 7 deaths in the remaining 18 cases. Although
209
operation seems to have been withheld in some of the fatalcases because the patients were in precarious condition, a review of the recordssuggests that some of these lives might have been saved by a more aggressivepolicy of surgical intervention.
Incisional sepsis-Minor degrees of incisional sepsis, which were seldomrecorded, were not uncommon. Major sepsis was infrequent, the 32 recorded cases representingapproximately 1 percent of the whole series.
The best method of preventing wound infection was found to be the routinedrainage of subcutaneous tissues. It became the established practice toaccomplish this by the following technique: A thin strip of fine-mesh gauze or asmall bandage was laid longitudinally in the incision, before the skin wasclosed, so as to keep the cutaneous margins and subcutaneous fat separated. Theskin sutures were then tied loosely across it. The gauze was left in place untilthe second or third day. It was the clinical impression that drainage ofsubcutaneous fat, which was particularly vulnerable to infection, materiallyexpedited healing by first intention.
Because the incisions used for abdominal wounds were almost alwayspotentially infected, it was necessary to change the dressings and inspect thewounds more frequently than is customary in civilian practice. Surgical drainagewas instituted promptly whenever suppuration developed.
Wound dehiscence-The incidence of wound dehiscence was higher in thisseries than in civilian practice, for reasons outlined elsewhere (p. 196).Prompt resuture was the treatment of choice.
Gastrointestinal fistula-The 23 instances of gastrointestinal fistularecorded in this series represent a relatively small incidence in view of thelarge number of perforations of hollow viscera. It is likely that additionalfistulas were observed after the patients were evacuated, since this isfrequently a somewhat delayed development.
Gastrointestinal fistulas, while they did not represent immediateemergencies, were considered an indication for priority of evacuation to fixedhospitals, where more adequate therapeutic facilities were available than infield and evacuation hospitals. The basis of treatment was to provide thepatient with maximum nutritional support and to employ gastrointestinaldecompression whenever this measure was indicated.
Intraperitoneal abscess-The 15 subhepatic or subphrenic abscesses andthe 9 pelvic abscesses recorded in this series represent an incidence ofslightly less than 1 percent. This is not excessive, in view of the type ofwounds treated and the fact that this is another complication which probablypresented itself in a number of other cases after evacuation. The treatment wassurgical drainage.
Anaerobic infections-Clostridial infections of the abdomen wereencountered after operation only eight times. Death occurred in every instance.In five cases, the retroperitoneal space was infected. In the other three cases,the process involved the abdominal wall, and en bloc excision of the entirerectus abdominis muscle on one side was followed by wound dehiscence.
210
Treatment always included massive doses of polyvalent anti-gas-gangrene sera,in addition to penicillin and sulfadiazine in large doses.
Secondary hemorrhage-Only eight instances of secondary hemorrhage wererecorded. There is no doubt that moderate bleeding, such as might occur afterremoval of a pack from the liver, took place in many other cases but was notrecorded. In 2 of the 8 cases, the bleeding was from remote vessels, in 1instance an intercostal artery and in 1 instance the femoral artery. In theother 6 cases, the hemorrhage was from sutured gastric wounds (p. 230).
These six cases comprise the only recorded instances of gastrointestinalbleeding, which is probably to be explained by the fact that, although a verylarge number of gastrointestinal repairs were carried out, the operations werechiefly on the small intestine, in which suture lines are less likely to bleedthan suture lines in the colon. The policy of exteriorization of wounds of thecolon thus not only protected the patient against necrosis and leakage butapparently reduced the risk of postoperative hemorrhage as well.
Thromboembolism-The 22 instances of proved or suspected pulmonaryembolism, 21 of which were fatal, must be considered as merely an approximation,since autopsies were not performed in at least 40 percent of all deaths.Nonfatal emboli undoubtedly occurred, but they were either not diagnosed or notrecorded. Nonfatal thrombophlebitis and phlebothrombosis were recorded only fivetimes but were probably much more frequent.
In the majority of cases in which emboli were shown to have originated in theveins of the leg, there had also been wounds in the involved extremities. In oneinstance, the autopsy showed the source of the embolus to be the right externaland common iliac veins. This particular patient had sustained a wound of thepelvis, with perforation of the bladder and extensive extraperitoneal trauma,and the surgeon, because of the evidence of severe damage near the great vesselsin the right side of the pelvis, debated at operation the propriety ofprophylactic interruption of the right common iliac vein, even though it had notbeen directly traumatized. Unfortunately, he did not act affirmatively on hispremonition of trouble, and the man lost his life.
Prophylactic or therapeutic interruption of the femoral veins was employedonly occasionally in this series of abdominal injuries. This prophylacticprocedure should probably have been employed more frequently, in view of theapparent tendency for an embolism to arise in a wounded extremity, especiallywhen an associated abdominal wound increased the hazard. This was particularlytrue when, as was usually the case, the limb had to be immobilized in plaster ofparis, which not only compounded the risk of thromboembolism but also precludedfrequent examinations for the detection of postoperative phlebothrombosis. Thecase history just related also indicates the importance of giving seriousconsideration to ligation of the great veins of the pelvis when trauma toadjacent tissues has occurred.
Since the anticoagulant drugs (heparin, Dicumarol(bishydroxycoumarin, U. S. P.)) were not available in field hospitals, nocomment can be made on their efficacy in the treatment of thromboembolism. Itcan be said, however,
211
that in many abdominal injuries their employment would beboth difficult and dangerous. They would certainly be contraindicated in woundsof the liver or in associated wounds of the soft tissues, lung, or head, becauseof the risk of hemorrhage. Even in future wars, it is doubtful that heparincould be made available in adequate quantities and equally doubtful thatadequate facilities could be provided for the use of Dicumarol under fieldconditions. For these reasons, if needless fatalities from pulmonary embolismare to be prevented, it would seem that the military surgeon of the future mustbe well versed in the indications for, and techniques of, prophylacticinterruption of the femoral veins.
Anuria.-Anuria (posttraumatic renal failure, lower nephronnephrosis, pigment nephropathy) was recorded in 36 cases in this series, 35 ofwhich were fatal. Very little, was known about this complication early in thewar, and there is no doubt that the diagnosis was overlooked at that time. Withincreasing experience, however, surgeons became aware of its frequency and itsdangers and were even able to predict that it might develop in certain cases.
There were four major warning signs of impending renal failure:
1. While it was insidious in onset and usually could not bediagnosed positively until the third or fourth postoperative day, it wasparticularly likely to occur in the most severely wounded men, particularly whenshock had been profound and of long duration. Earlier than the third day, renalfailure could not be distinguished from the physiologic oliguria whichaccompanied shock and which might persist for 48 hours after operation.
2. Renal failure was a possibility whenever diuresis did not occur at the endof this time and the patient, if he was out of shock, continued to be oliguric.It was thus of prime importance to measure the daily volume of urine accurately.Once shock had been controlled and the blood pressure was tending to a normallevel, a 24-hour output of urine of less than 700 cc. was an ominous sign.
3. Hypertension was the next warning sign. With developingrenal failure, there was usually a slow increase in the blood pressure, whichsometimes climbed to 180/120 mm. Hg.
4. Azotemia was another sign of impending trouble. It was not uncommonimmediately after operation, when the nonprotein nitrogen level of the bloodmight reach 80 mg. percent. Usually, however, the value returned to normal on orabout the second postoperative day, when diuresis occurred. Persistent orincreasing azotemia after this time usually meant impending anuria. Thenonprotein nitrogen of the blood was readily determined under field conditionsby the copper sulfate method.
The treatment of renal failure was chiefly under the direction of the Boardfor the Study of the Severely Wounded, as a research project, and extensivedetails are contained in the report of that group.1 Results were disheartening.While management was not standardized, the general plan was to restrict
1Medical Department, UnitedStates Army, Surgery in World War II. The Physiologic Effects of Wounds.Washington: U. S. Government Printing Office, 1952.
212
fluid allowances to 1,000 cc. or less for the 24-hour period, this being theminimum required to compensate for insensible fluid losses. Salt was withheldentirely. The purpose of these restrictions was to prevent the pulmonary edemawhich was common in these cases. The urinary output was accurately measured bymeans of an indwelling catheter. Diuretics, including ethanol (5 percent) byvein, were tried but proved of little help. Alkalinization of the urine withsalts of lactic acid was also tried but was later abandoned as useless anddangerous; renal block prevented the desired effect, and serious alkalosisresulted. Peritoneal lavage was equally ineffective.
Miscellaneous complications-Other complications, most of which werefatal, were recorded only occasionally. According to the clinical charts,urinary fistula occurred 6 times; fat embolism, 4 times; acute gastricdilatation, 3 times; and vesicorectal fistula, acute noncontagious parotitis, acutenonspecific orchitis, encephalomalacia following ligation of the common carotidartery, meningitis secondary to spinal cord injury, cerebral infarct,anaphylactic shock following the use of intravenous protein hydrolysate,cachexia associated with ileostomy, and air embolism, in 1 case each.
